def cmodel():
model = Sequential()
# Dense 1
model.add(Dense(1024, input_shape=(90, )))
model.add(Activation('relu'))
model.add(Dropout(0.1))
model.add(Dense(1024))
model.add(BN())
model.add(GN(0.1))
model.add(Activation('relu'))
model.add(Dense(512))
model.add(BN())
model.add(GN(0.1))
model.add(Activation('relu'))
model.add(Dense(num_classes))
model.add(Activation('softmax'))
model.summary()
sgd = SGD(lr=0.0, momentum=0.9, decay=0.0, nesterov=False)
model.compile(loss='categorical_crossentropy',
optimizer=sgd,
metrics=['accuracy'])
return model
def setModel(num_classes, shape):
model = Sequential()
model.add(Conv2D(32, (3, 3), padding='same', input_shape=shape))
model.add(BN())
model.add(GN(0.3))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Conv2D(64, (3, 3), padding='same'))
model.add(BN())
model.add(GN(0.3))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Conv2D(512, (3, 3), padding='same'))
model.add(BN())
model.add(GN(0.3))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Flatten())
model.add(Dense(512))
model.add(BN())
model.add(GN(0.3))
model.add(Activation('relu'))
model.add(Dense(num_classes))
model.add(Activation('softmax'))
sgd = SGD(lr=0.01, momentum=0.9, decay=0.0, nesterov=False)
model.compile(loss='categorical_crossentropy',
optimizer=sgd,
metrics=['accuracy'])
return model
def model_keras():
# crea el modeloTrue
model = Sequential()
# primera capa
model.add(Dense(128, input_shape=(87, )))
model.add(BN())
model.add(GN(0.3))
model.add(Activation('relu'))
# segunda capa
model.add(Dense(256))
model.add(BN())
model.add(GN(0.3))
model.add(Activation('relu'))
# tercera capa
model.add(Dense(256))
model.add(BN())
model.add(GN(0.3))
model.add(Activation('relu'))
# capa de salida
model.add(Dense(1))
model.add(GN(0.3))
model.add(Activation('relu'))
# compilar el modelo
model.compile(loss='mean_squared_error', optimizer='adam')
return model
def basic_model(): # 1024 512
model = Sequential()
model.add(Dense(2048, input_shape=(90, )))
model.add(Activation('relu'))
model.add(Dropout(0.1))
model.add(Dense(1024))
model.add(BN())
model.add(GN(0.1))
model.add(Activation('relu'))
model.add(Dense(512))
model.add(BN())
model.add(GN(0.1))
model.add(Activation('relu'))
model.add(Dense(1))
model.add(Activation('relu'))
model.summary()
adam = keras.optimizers.Adam(lr=0.01,
beta_1=0.9,
beta_2=0.999,
epsilon=None,
decay=0.0,
amsgrad=False)
model.compile(loss='mape', optimizer=adam, metrics=['mse'])
return model
def setModel(num_classes):
model = Sequential()
model.add(Dense(512, input_shape=(784,)))
model.add(BN())
model.add(GN(0.3))
model.add(Activation('relu'))
model.add(Dense(512))
model.add(BN())
model.add(GN(0.3))
model.add(Activation('relu'))
model.add(Dense(num_classes, activation='softmax'))
sgd=SGD(lr=0.01, decay=1e-6, momentum=0.9)
model.compile(loss='categorical_crossentropy',optimizer=sgd,metrics=['accuracy'])
return model
def CBGN(model, filters, lname, ishape=0):
if (ishape != 0):
model.add(Conv2D(filters, (3, 3), padding='same', input_shape=ishape))
else:
model.add(Conv2D(filters, (3, 3), padding='same'))
model.add(BN())
model.add(GN(0.3))
model.add(Activation('relu'))
model.add(Conv2D(filters, (3, 3), padding='same'))
model.add(BN())
model.add(GN(0.3))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2), name=lname))
return model
def main():
batch_size = 1024
num_classes = 2
epochs = 500
#Load created npy arrays with vectorized images
train_x = np.load("train_X_FOURTH_iter5.npy", allow_pickle=True)
#Load labels
train_y_orig = np.load("train_Y_FOURTH_iter5.npy", allow_pickle=True)
train_x = train_x.astype("float32")
train_y = train_y_orig.astype("float32")
#Normalize data
train_x /= 255
#One-hot vector of labels
num_classes = 2
train_y = keras.utils.to_categorical(train_y, num_classes)
#Data augmentation on the fly
datagen = ImageDataGenerator(
width_shift_range=0.1, #horizontal shift
height_shift_range=0.1, #Vertical shift
rotation_range=10, #Image rotation [-10, +10] degrees
horizontal_flip=True #Flip the image
)
#Train the datagenerator over the train set
datagen.fit(train_x)
"""
###### CONSTRUCTION OF MODEL ######
"""
model = Sequential()
model.add(Conv2D(8, (3, 3), padding='same', strides=1, input_shape=train_x.shape[1:]))
model.add(Activation("relu"))
model.add(Conv2D(16, (3, 3), padding='valid', strides=2))
model.add(Activation("relu"))
model.add(Conv2D(32, (3, 3), padding='valid', strides=1))
model.add(Activation("relu"))
model.add(Conv2D(16, (3, 3), padding='valid', strides=1))
model.add(Activation("relu"))
model.add(Conv2D(8, (3, 3), padding='valid', strides=1))
model.add(Activation("relu"))
model.add(Flatten())
model.add(GN(0.3))
model.add(Dense(512, activation="relu"))
model.add(Dense(num_classes, activation="softmax"))
model.summary()
opt = SGD(lr=0.001, decay=1e-6, momentum=0.9)
#Only if fine tunning a previously trained model. If so, reduce learning rate above.
#model = load_model("model_faces_20_iter6_dev.hdf5")
model.compile(loss='categorical_crossentropy',optimizer=opt,metrics=['accuracy'])
save_model = keras.callbacks.ModelCheckpoint( "model_faces_20_iter6_dev_DA.hdf5", monitor='accuracy', verbose=1, save_best_only=True, save_weights_only=False, mode='max')
history=model.fit_generator(datagen.flow(train_x, train_y,batch_size),
steps_per_epoch=len(train_x) / batch_size,
epochs=epochs,
callbacks=[save_model],
verbose=1)
height_shift_range=0.2,
horizontal_flip=False,
vertical_flip=True)
datagen.fit(x_train)
"""Create Model for training"""
shape = x_train.shape[1:]
model = Sequential()
model.add(BN(input_shape=shape))
model.add(Dropout(0.1))
model.add(Conv2D(32, (3, 3), padding='same', kernel_initializer=glorot_uniform(seed=0)))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(BN())
model.add(GN(0.3))
model.add(Flatten())
model.add(Dense(256, kernel_initializer=glorot_uniform(seed=0)))
model.add(Activation('relu'))
model.add(BN())
model.add(Dropout(0.1))
model.add(Dense(128, kernel_initializer=glorot_uniform(seed=0)))
model.add(Activation('relu'))
model.add(BN())
model.add(Dropout(0.1))
model.add(Dense(1))
model.add(Activation('sigmoid'))
# convert class vectors to binary class matrices
y_train = keras.utils.to_categorical(y_train, num_classes)
y_test = keras.utils.to_categorical(y_test, num_classes)
## Data Augmentation with an ImageGenerator
datagen = ImageDataGenerator(width_shift_range=0.1,
height_shift_range=0.1,
horizontal_flip=False)
#Define the NN topology, a sequential model with 2 hidden layers
model = Sequential()
# noise in input
model.add(Reshape(target_shape=(784, ), input_shape=(28, 28, 1)))
model.add(GN(0.1))
model.add(Dense(1024, input_shape=(784, )))
model.add(BN())
model.add(GN(0.1))
model.add(Activation('relu'))
model.add(Dropout(0.02))
model.add(Dense(1024, input_shape=(784, )))
model.add(BN())
model.add(GN(0.1))
model.add(Activation('relu'))
model.add(Dropout(0.02))
model.add(Dense(1024))
model.add(BN())
rotation_range=20,
zoom_range=[1.0, 1.2],
horizontal_flip=True)
## DEF CNN TOPOLOGY 1
model1 = Sequential()
model1 = CBGN(model1, 32, 'conv_model1_1', x_train.shape[1:])
model1 = CBGN(model1, 64, 'conv_model1_2')
model1 = CBGN(model1, 128, 'conv_model1_3')
model1 = CBGN(model1, 256, 'conv_model1_4')
model1 = CBGN(model1, 512, 'conv_model1_5')
model1.add(Flatten())
model1.add(Dense(512))
model1.add(BN())
model1.add(GN(0.3))
model1.add(Activation('relu'))
model1.add(Dropout(0.5))
model1.add(Dense(num_classes))
model1.add(Activation('softmax'))
model1.summary()
## OPTIM AND COMPILE
opt = SGD(lr=0.1, decay=1e-6)
model1.compile(loss='categorical_crossentropy',
optimizer=opt,
metrics=['accuracy'])
# DEFINE A LEARNING RATE SCHEDULER
'n_channels': 1,
'n_classes': 2,
'shuffle': True,
'dir_faces': './face_images/',
'dir_no_faces': './no_face_images/'
}
training_generator = DataGenerator(f_train, nf_train, **params)
test_generator = DataGenerator(f_test, nf_test, **params)
validation_generator = DataGenerator(f_val, nf_val, **params)
print("Building Model")
# Start building our model
model = Sequential()
model.add(BN(input_shape=(32, 32, 1)))
model.add(GN(gn))
model.add(Conv2D(32, (3, 3), padding='same'))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(BN())
model.add(GN(gn))
model.add(Conv2D(64, (3, 3), padding='same'))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(BN())
model.add(GN(gn * 1.5))
model.add(Conv2D(128, (3, 3), padding='same'))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))