Question

我想在学习过程中向cnn中间层的输出层添加一个常数矩阵，然后将其发送到下一层。我将代码放在这里并使用添加功能，但会产生错误。我该怎么办？使用添加是否是真正的解决方案？

from keras.layers import Input, Concatenate, GaussianNoise
from keras.layers import Conv2D
from keras.models import Model
from keras.datasets import mnist
from keras.callbacks import TensorBoard
from keras import backend as K
from keras import layers
import matplotlib.pyplot as plt
import tensorflow as tf
import keras as Kr
import numpy as np

w_main = np.random.randint(2,size=(1,4,4,1))
w_main=w_main.astype(np.float32)
w_expand=np.zeros((1,28,28,1),dtype='float32')
w_expand[:,0:4,0:4]=w_main
w_expand.reshape(1,28,28,1)
#-----------------------encoder------------------------------------------------
#------------------------------------------------------------------------------

image = Input((28, 28, 1))
conv1 = Conv2D(8, (5, 5), activation='relu', padding='same')(image)
conv2 = Conv2D(4, (3, 3), activation='relu', padding='same')(conv1)
conv3 = Conv2D(2, (3, 3), activation='relu', padding='same')(conv2)
encoded =  Conv2D(1, (3, 3), activation='relu', padding='same')(conv3)

encoder=Model(inputs=image, outputs=encoded)
encoder.summary()
#-----------------------adding w---------------------------------------
encoded_merged=Kr.layers.Add(encoded,w_expand)

#-----------------------decoder------------------------------------------------
#------------------------------------------------------------------------------

#encoded_merged = Input((28, 28, 2))
x = Conv2D(2, (5, 5), activation='relu', padding='same')(encoded_merged)
x = Conv2D(4, (3, 3), activation='relu', padding='same')(x)
x = Conv2D(8, (3, 3), activation='relu',padding='same')(x)
decoded = Conv2D(1, (3, 3), activation='sigmoid', padding='same', name='decoder_output')(x) 

decoder=Model(inputs=encoded_merged, outputs=decoded)
decoder.summary()

产生的错误是：

TypeError： init （）接受1个位置参数，但给出了3个我很急。请帮助我。

Answer 1

您使用错误的图层方式，这是正确的方法：

encoded_merged=Kr.layers.Add()([encoded,w_expand])

Answer 2

您需要将常量包装到将返回张量的Layer中，当前您有numpy数组，无法将其添加到张量：

add_const = Kr.layers.Lambda(lambda x: x + Kr.backend.constant(w_expand))

并与要添加到的图层一起使用：

encoded_merged = add_const(encoded)

向cnn的中间层添加常数

2 个答案: