在 Keras 中创建 CNN 模型时出现属性错误

Question

用于构建模型的函数：

def CancerModel(input_shape):   

    X_input = Input(input_shape)

    X = ZeroPadding2D((2, 2))(X_input)

    X = Conv2D(8, (5, 5), strides = (1, 1), name = 'conv')(X)
    X = BatchNormalization(axis = 3, name = 'bn1')(X)
    X = Activation('relu')(X)

    X = MaxPooling2D((2, 2), name='max_pool')(X)

    X = Conv2D(16, (5, 5), strides = (1, 1), name = 'conv2')(X)
    X = BatchNormalization(axis = 3, name = 'bn2')(X)
    X = Activation('relu')(X)

    X = MaxPooling2D((2, 2), name='max_pool2')(X)

    X = Flatten()(X)
    X = Dense(120, activation='relu', name='fc1')(X)
    X = Dense(84, activation='relu', name='fc2')(X)
    X = Dense(7, activation='softmax', name='output')(X)

    model = Model(inputs = X_input, outputs = output, name='CancerModel')

    return model

尝试使用以下方法创建模型：

cancerModel = CancerModel(X_train.shape[1:])

但是，我收到一条错误消息，指出无法设置该属性。我还附上了我收到的错误的屏幕截图。任何帮助将不胜感激。

Screenshot of bug

Answer 1

我发现你的代码没有错误。可能是你的 training data shape 或 Keras 版本问题（我的 Keras 版本 2.4.3）。

def CancerModel(input_shape):   

    X_input = Input(input_shape)

    X = ZeroPadding2D((2, 2))(X_input)

    X = Conv2D(8, (5, 5), strides = (1, 1), name = 'conv')(X)
    X = BatchNormalization(axis = 3, name = 'bn1')(X)
    X = Activation('relu')(X)

    X = MaxPooling2D((2, 2), name='max_pool')(X)

    X = Conv2D(16, (5, 5), strides = (1, 1), name = 'conv2')(X)
    X = BatchNormalization(axis = 3, name = 'bn2')(X)
    X = Activation('relu')(X)

    X = MaxPooling2D((2, 2), name='max_pool2')(X)

    X = Flatten()(X)
    X = Dense(120, activation='relu', name='fc1')(X)
    X = Dense(84, activation='relu', name='fc2')(X)
    X = Dense(7, activation='softmax', name='output')(X)

    model = Model(inputs = X_input, outputs = X, name='CancerModel')

    return model


CancerModel((224,224,3)).summary() #It works fine

使用

也能正常工作

CancerModel(np.ones((5,120,120,3)).shape[1:]).summary()

在 Keras 中创建 CNN 模型时出现属性错误

Attribute error while creating CNN model in Keras

attributeerror

conv-neural-network

keras