1

连体网络

我正在尝试实现连体神经网络，并运行此代码

def build_network(conv_model):

    input_shape = (105, 105, 1)
    input1 = Input(input_shape)
    input2 = Input(input_shape)

    model = conv_model(input_shape)

    model_output_left = model(input1)
    model_output_right = model(input2)

    def l1_distance(x): 
        return K.abs(x[0] - x[1])

    def l1_distance_shape(x): 
        print(x)
        return x[0]



    # merged_model = Merge([model1, model2], mode=l1_distance, output_shape=lambda x: x[0])
    merged_model = merge([model_output_left, model_output_right], mode=l1_distance, output_shape=l1_distance_shape)
    output = Dense(1, activation='sigmoid')(merged_model)
    siamese_model = Model([input1, input2], output)
    return siamese_model

然后我跑去制作模型

siamese_model1 = build_network(conv_model)
siamese_model1.compile(loss='binary_crossentropy', optimizer=Adam(0.00006), metrics=['acc'])
siamese_model1.summary()

然后我得到错误

---------------------------------------------------------------------------

TypeError                                 Traceback (most recent call last)

<ipython-input-28-7405baecdb7f> in <module>()
----> 1 siamese_model1 = build_network(conv_model)
      2 siamese_model1.compile(loss='binary_crossentropy', optimizer=Adam(0.00006), metrics=['acc'])
      3 siamese_model1.summary()

<ipython-input-27-294ae7b24fbc> in build_network(conv_model)
     20 
     21     # merged_model = Merge([model1, model2], mode=l1_distance, output_shape=lambda x: x[0])
---> 22     merged_model = merge([model_output_left, model_output_right], mode=l1_distance, output_shape=l1_distance_shape)
     23     output = Dense(1, activation='sigmoid')(merged_model)
     24     siamese_model = Model([input1, input2], output)

TypeError: 'module' object is not callable

我有人帮我解决这个问题吗？或评论如何解决？

4

1 回答 1

0

您可以merge用Lambda图层替换。

from keras.layers import Lambda
merged_model = Lambda(l1_distance, output_shape=l1_distance_shape)([model_output_left, model_output_right])
于 2021-09-18T02:56:02.770 回答