Question

我试图在ResNet50上添加一个Flatten图层，一个Dense图层（relu）和一个Dense图层（softmax），用于在Win10.Here上使用Keras 2.0.2 Theano 0.9.0 py2.7进行多分类任务我的代码：

def create_model():
    base_model = ResNet50(include_top=False, weights=None,
                            input_tensor=None, input_shape=(3,224,224),
                            pooling=None)

    base_model.load_weights(weight_path+'/resnet50_weights_th_dim_ordering_th_kernels_notop.h5')
    x = base_model.output
    x = Flatten()(x)
    x = Dense(128,activation='relu',kernel_initializer='random_normal',
            kernel_regularizer=regularizers.l2(0.1),
            activity_regularizer=regularizers.l2(0.1))(x)

    x=Dropout(0.3)(x)
    y = Dense(8, activation='softmax')(x)
    model = Model(base_model.input, y)
    for layer in base_model.layers:
        layer.trainable = False
    model.compile(optimizer='adadelta',
    loss='categorical_crossentropy')
    return model

我设置了image_dim_ordering：

from keras import backend as K
K.set_image_dim_ordering('th')

这是我的Keras.json文件：

"backend": "theano", ``"image_data_format": "channels_first", ``"epsilon": 1e-07, ``"floatx": "float32"}

以下是错误消息：

ValueError: The shape of the input to "Flatten" is not fully defined (got (2048, None, None). Make sure to pass a complete "input_shape" or "batch_input_shape" argument to the first layer in your model.

Answer 1

你应该

将input_shape参数传递给第一层。这是一个形状元组（整数或无条目的元组，其中None表示可以预期任何正整数）。在input_shape中，不包括批量维度。

在您的情况下，第一层是Flatten（）图层。它应该像

your_input = Input(shape=output_shape_of_resnet)
x = Flatten(your_input)

至于将resnet50的输出提供给您自己的图层，请考虑定义一个包含您自己的图层和resnet的新模型，例如

 new_model = Sequential()
 new_model.add(resnet_model) #Of course you need the definition and weights of resnet
 resnet_model.trainable = False #I guess?
 new_model.add(your_own_layers_model)

Answer 2

如果输入图像的大小对于网络模型来说太小，我遇到了一些错误。如果图层的输出数据大小变为0，则会出现此错误。您可以使用model.summary()查看您的网络外观。这是model.summary()输出的示例：

Layer (type)                 Output Shape              Param #   
=================================================================
conv2d_78 (Conv2D)           (None, 16, 21, 21)        160       
_________________________________________________________________
max_pooling2d_62 (MaxPooling (None, 16, 5, 5)          0         
_________________________________________________________________
...
flatten_25 (Flatten)         (None, 32)                0         
_________________________________________________________________
dense_28 (Dense)             (None, 2)                 1026      
=================================================================
Total params: 31,970
Trainable params: 31,970
Non-trainable params: 0
_________________________________________________________________

尝试在ResNet50（notop）上添加Flatten图层并出错

2 个答案: