AttributeError: 'function' object 在构建 CNN 时没有属性 'summary'

Question

I am trying to build a CNN model. But I am encountering this error and cannot resolve it.

from keras.models import Sequential
from keras.layers import Conv2D,MaxPooling2D,Flatten,Dense
from keras.optimizers import Adam
def vgg():
    kernel=(3,3)
    pool=(2,2)
    model=Sequential()
    model.add(Conv2D(96,kernel,padding='same',input_shape=(32,32,1),activation='relu'))
    model.add(Conv2D(96,kernel,padding='same',activation='relu'))
    model.add(MaxPooling2D(pool,strides=2,padding='same'))

    model.add(Conv2D(128,kernel,padding='same',activation='relu'))
    model.add(Conv2D(128,kernel,padding='same',activation='relu'))
    model.add(MaxPooling2D(pool,strides=2,padding='same'))

    model.add(Conv2D(256,kernel,padding='same',activation='relu'))
    model.add(Conv2D(256,kernel,padding='same',activation='relu'))
    model.add(Conv2D(256,kernel,padding='same',activation='relu'))
    model.add(MaxPooling2D(pool,strides=2,padding='same'))

    model.add(Conv2D(512,kernel,padding='same',activation='relu'))
    model.add(Conv2D(512,kernel,padding='same',activation='relu'))
    model.add(Conv2D(512,kernel,padding='same',activation='relu'))
    model.add(MaxPooling2D(pool,strides=2,padding='same'))

    model.add(Conv2D(512,kernel,padding='same',activation='relu'))
    model.add(Conv2D(512,kernel,padding='same',activation='relu'))
    model.add(Conv2D(512,kernel,padding='same',activation='relu'))
    model.add(MaxPooling2D(pool,strides=2,padding='same'))

    model.add(Flatten())
    model.add(Dense(4096,activation='relu'))
    model.add(Dense(4096,activation='relu'))

    model.add(Dense(43,activation='softmax'))
    model.compile(Adam(lr=0.001),loss='categorical_crossentropy',metrics=['accuracy'])
    return vgg

model=vgg()
print(model.summary())

I could'nt find the solution to it. Is it due to the more no of filters or due to higher values in dense, cause in another model I have used lesser values and they are working fine

Answer 1

It's super simple, you need to return the model not the function itself. So, in vgg , you need to return model :

def vgg():
    ...
    ...
    return model #not vgg