def get_data_all(path_data, array_max):
data_all = Dataset_csv(path_data=path_data, max_value=array_max)
data_all.set_minibatch(data_all.total_inputs)
X_data, y_data = data_all.generate_batch()
return X_data, y_data, len(y_data)
def get_data_split(path_data, array_max, test_size=0.3):
data_all = Dataset_csv(path_data=path_data, max_value=array_max)
data_all.set_minibatch(data_all.total_inputs)
data, label = data_all.generate_batch()
X_train, X_test, y_train, y_test = model_selection.train_test_split(
data, label, test_size=test_size, random_state=42)
return X_train, X_test, y_train, y_test, len(y_train), len(y_test)
def get_data_all(path_train, path_test, array_max):
data_all = Dataset_csv(path_data=path_train, max_value=array_max)
data_all.set_minibatch(data_all.total_inputs)
X_train, y_train = data_all.generate_batch()
data_all = Dataset_csv(path_data=path_test, max_value=array_max)
data_all.set_minibatch(data_all.total_inputs)
X_test, y_test = data_all.generate_batch()
return X_train, X_test, y_train, y_test, len(y_train), len(y_test)
path_load_weight0 = '../weight/tlconvolae_dual_' + OPC + '_1.npy'
path_save_weight0 = '../weight/tlconvolae_class0_' + OPC + '_1.npy'
path_load_weight1 = '../weight/tlconvolae_dual_' + OPC + '_1.npy'
path_save_weight1 = '../weight/tlconvolae_class1_' + OPC + '_1.npy'
if __name__ == '__main__':
mini_batch_train = 34
mini_batch_test = 30
epoch = 10
learning_rate = 0.0001
# Datos de valor maximo
data_normal = Dataset_csv(
path_data=[path_data_train_dual[0], path_data_test_dual[0]],
random=False)
Damax = data_normal.amax
del data_normal
# utils.generate_max_csvData([path_data_train_dual[0], path_data_test_dual[0]], path+'maximo.csv', has_label=True)
# Damax = utils.load_max_csvData(path+'maximo.csv')
# -------------------------------------------------------------------
# ENTRENAMOS EL AUTOENCODER CON AMBAS CLASES - GENERAMOS UN PESO BASE
# -------------------------------------------------------------------
print('ConvAE TRAIN DUAL')
print('-----------------')
data_train = Dataset_csv(path_data=path_data_train_dual,
minibatch=mini_batch_train,
# ..................................................................
if __name__ == '__main__':
path_load_weight = None
path_save_weight = '../weight/saveAE_1.npy'
mini_batch_train = 20
mini_batch_test = 25
epoch = 10
learning_rate = 0.0001
noise_level = 0
# Datos de media y valor maximo
data_normal = Dataset_csv(
path_data=[path_data_train[0], path_data_test[0]], random=False)
Damax = data_normal.amax
# utils.generate_max_csvData([path_data_train[0], path_data_test[0]], path+'maximo.csv', has_label=True)
# Damax = utils.load_max_csvData(path+'maximo.csv')
# Load data train
data_train = Dataset_csv(path_data=path_data_train,
minibatch=mini_batch_train,
max_value=Damax)
# Load data test
data_test = Dataset_csv(path_data=path_data_test,
minibatch=mini_batch_test,
max_value=Damax,
random=False)
# data_test = Dataset_csv(path_data=path_data_train, minibatch=mini_batch_train, max_value=Damax, random=False)
# data_normal = Dataset_csv(path_data=[path_data_train_all[0], path_data_test_all[0]], random=False)
# Damax = data_normal.amax
# del data_normal
# utils.generate_max_csvData([path_data_train_all[0], path_data_test_all[0]], path+'maximo.csv', has_label=True)
Damax = utils.load_max_csvData(path + 'maximo.csv')
c = tf.ConfigProto()
c.gpu_options.visible_device_list = "1,2"
print('SEARCH SAMPLES')
print('--------------')
data = Dataset_csv(path_data=path_data_test_all,
minibatch=1,
max_value=Damax,
restrict=False,
random=False)
with tf.device('/cpu:0'):
with tf.Session(config=c) as sess:
aencoder = AE.ae_multiClass(session=sess,
npy_weight_paths=path_weight_ae,
num_class=num_class)
aencoder.build(dim_input=dim_input, layers=layers)
for i in range(data.total_batchs_complete):
x, label = data.generate_batch()
res = aencoder.search_sample(sample=x)
data.next_batch_test()
print(res, label)
#data_normal = Dataset_csv(path_data=[path_data_train_all[0], path_data_test_all[0]], random=False)
#Damax = data_normal.amax
#del data_normal
# utils.generate_max_csvData([path_data_train_all[0], path_data_test_all[0]], path+'maximo.csv', has_label=True)
Damax = utils.load_max_csvData(path+'maximo.csv')
c = tf.ConfigProto()
c.gpu_options.visible_device_list = "1,2"
# -------------------------------------------------------------------
# ENTRENAMOS EL AUTOENCODER CON AMBAS CLASES - GENERAMOS UN PESO BASE
# -------------------------------------------------------------------
print('AE TRAIN ALL')
print('------------')
data_train = Dataset_csv(path_data=path_data_train_all, minibatch=mini_batch_train, max_value=Damax)
print('Load data train...')
data_test = Dataset_csv(path_data=path_data_test_all, minibatch=mini_batch_test, max_value=Damax, random=False)
print('Load data test...')
with tf.Session(config=c) as sess:
x_batch = tf.placeholder(tf.float32, [None, dim_input])
AEncode = AE.AEncoder(path_load_weight_all, learning_rate=learning_rate_all)
AEncode.build(x_batch, layers)
sess.run(tf.global_variables_initializer())
print('Original Cost: ', test_model(AEncode, sess, data_test))
train_model(AEncode, sess, data_train, objDatatest=data_test, epoch=epoch_all)
for i in range(lenght):
f.write(','.join(map(str, X[i])) + '\n')
f.close()
if __name__ == '__main__':
c = tf.ConfigProto()
c.gpu_options.visible_device_list = "1,2"
print('CNN + AE + LSH')
print('--------------')
# data_train = Dataset_csv(path_data=path_data_train_all, minibatch=30, max_value=1, restrict=False, random=True)
data = Dataset_csv(path_data=path_data_test_all,
minibatch=30,
max_value=1,
restrict=False,
random=False)
with tf.device('/cpu:0'):
with tf.Session() as sess:
calsh = CAL.cnn_ae_lsh(session=sess,
npy_convol_path=path_w_cnn,
npy_ae_path=path_w_ae_all,
npy_ae_class_paths=path_w_ae_class,
normal_max_path=path_normalization_max,
num_class=num_class,
k_classes=1)
calsh.build(dim_input=dim_input, layers=layers)
path_data_train = [path + 'SKINfeatures' + OPC + '_Train.csv']
path_data_test = [path + 'SKINfeatures' + OPC + '_Test.csv']
path_load_weight = None
path_save_weight = '../weight/tlmlp_' + OPC + '_1.npy'
if __name__ == '__main__':
mini_batch_train = 20
mini_batch_test = 30
learning_rate = 0.0001
epoch = 5
num_class = 2
# GENERATE DATA
# Datos de media y valor maximo
data_normal = Dataset_csv(
path_data=[path_data_train[0], path_data_test[0]], random=False)
Damax = data_normal.amax
# utils.generate_max_csvData([path_data_train[0], path_data_test[0]], path+'maximo.csv', has_label=True)
# Damax = utils.load_max_csvData(path+'maximo.csv')
# Load data train
data_train = Dataset_csv(path_data=path_data_train,
minibatch=mini_batch_train,
max_value=Damax,
restrict=True)
# Load data test
data_test = Dataset_csv(path_data=path_data_test,
minibatch=mini_batch_test,
max_value=Damax,
random=False)
accuracy = 0
# Damax = data_normal.amax
# del data_normal
# utils.generate_max_csvData([path_data_train_all[0], path_data_test_all[0]], path+'maximo.csv', has_label=True)
# Damax = utils.load_max_csvData(path + 'maximo.csv')
c = tf.ConfigProto()
c.gpu_options.visible_device_list = "1,2"
print('CNN + AE + LSH')
print('--------------')
#data_train = Dataset_csv(path_data=path_data_train_all, minibatch=30, max_value=1, restrict=False, random=True)
data = Dataset_csv(path_data=path_data_test_all,
minibatch=30,
max_value=1,
restrict=False,
random=False)
with tf.device('/cpu:0'):
with tf.Session() as sess:
calsh = CAL.cnn_ae_lsh(session=sess,
npy_convol_path=path_w_cnn,
npy_ae_path=path_w_ae_all,
npy_ae_class_paths=path_w_ae_class,
normal_max_path=path_normalization_max,
num_class=num_class,
k_classes=2,
threshold=0.0002)
calsh.build(dim_input=dim_input, layers=layers)
if __name__ == '__main__':
mini_batch_train = 35
mini_batch_test = 5
epoch = 50
learning_rate = 0.00005
l_hidden = 16
ratio_diff = 0.05
pca = True
Damax = utils.load_max_csvData(path_maximo)
# data_train = Dataset_csv(path_data=path_data_train, minibatch=mini_batch_train, max_value=Damax, restrict=False)
data_test = Dataset_csv(path_data=path_data_test,
minibatch=mini_batch_test,
max_value=Damax,
restrict=False,
random=False)
Xmatrix = genfromtxt(path_data_train[0], delimiter=',')
shape = np.shape(Xmatrix)
Xmatrix = Xmatrix[:, :shape[1] - 1]
XFractal = getfractal(path, path_data_test[0].split('/')[-1], Xmatrix)
print("fractal dimension of X:", XFractal)
c = tf.ConfigProto()
c.gpu_options.visible_device_list = "0"
with tf.Session() as sess:
x_batch = tf.placeholder(tf.float32, [None, dim_input])
if ep % 5 == 0:
cost_tot, cost_prom = test_model(net, sess_train, objData)
print(' Epoch', ep, ': ', cost_tot, ' / ', cost_prom)
if __name__ == '__main__':
epoch = 21
learning_rate = 0.00008
for opc in range(3,4):
path_data_train_csv, path_data_test_csv, path_max_csv, name, dims, method, origDim = path_datasets(opc)
Damax = utils.load_max_csvData(path_max_csv)
data_train = Dataset_csv(path_data=path_data_train_csv, minibatch=35, max_value=Damax, random=False)
# data_test = Dataset_csv(path_data=path_data_test_csv, minibatch=35, max_value=Damax, restrict=False, random=False)
print('[', name, ']')
for xdim in dims:
print(' Dim:', xdim)
pathFile = xpath + name + '/'
with tf.Session() as sess:
weight = xpath + name + '/' + 'weight-' + str(xdim) + '.npy'
layers = [[int(origDim / 2), 'relu'], [xdim, 'relu']]
x_batch = tf.placeholder(tf.float32, [None, origDim])
ae = AE.AEncoder(weight, learning_rate=learning_rate)
ae.build(x_batch, layers)
