AttributeError：“模型”对象没有属性“名称”

Question

我是Keras新手，当我遇到这个错误时，我正试图使用Python 3.6构建一个text-classification CNN模型：

AttributeError: 'Model' object has no attribute 'name'

这是我写的代码：

print("\nCreating Model...")
x1 = Input(shape=(seq_len1, 100), name='x1')
x2 = Input(shape=(seq_len2, 100), name='x2')
x1 = Reshape((seq_len1, embedding_dim, 1))(x1)
x2 = Reshape((seq_len2, embedding_dim, 1))(x2)

conv_0 = Conv2D(num_filters, kernel_size=(filter_sizes[0], 1), padding='valid', kernel_initializer='normal', activation='relu')
conv_1 = Conv2D(num_filters, kernel_size=(filter_sizes[1], 1), padding='valid', kernel_initializer='normal', activation='relu')
conv_2 = Conv2D(num_filters, kernel_size=(filter_sizes[2], 1), padding='valid', kernel_initializer='normal', activation='relu')

maxpool = MaxPool2D(pool_size=(2, 1), strides=(1,1), padding='valid')

output1 = conv_0(x1)
output1 = maxpool(output1)
output1 = conv_1(output1)
output1 = maxpool(output1)
output1 = conv_2(output1)
output1 = maxpool(output1)
.
.
# Same for output2
.
concatenated_tensor = Concatenate(axis=1)([output1, output2])
flatten = Flatten()(concatenated_tensor)
#dropout = Dropout(drop)(flatten)
output = Dense(units=1024, activation='relu')(flatten)
output = Dense(units=1024, activation='relu')(output)
output = Dense(units=1, activation='softmax')(output)

# this creates a model that includes
model = Model(inputs=[x1, x2], outputs=[output])

在最后一行中会遇到错误。请帮我解决这个问题

编辑:

Traceback (most recent call last):
  File "model.py", line 91, in <module>
    model = Model(inputs=[x1, x2], outputs=[out])
  File "/../../anaconda3/lib/python3.6/site-packages/keras/legacy/interfaces.py", line 91, in wrapper
    return func(*args, **kwargs)
  File "/../../anaconda3/lib/python3.6/site-packages/keras/engine/network.py", line 91, in __init__
    self._init_graph_network(*args, **kwargs)
  File "/../../anaconda3/lib/python3.6/site-packages/keras/engine/network.py", line 183, in _init_graph_network
    'The tensor that caused the issue was: ' +
AttributeError: 'Model' object has no attribute 'name'

Answer 1

x1和x2指向输入的重塑层，而不是输入层本身。

Answer 2

当我使用自定义的"swish“激活函数时，也会出现相同的错误。我可以用下面的代码解析相同的

class Swish(Activation):
    def __init__(self, activation, **kwargs):
        super(Swish, self).__init__(activation, **kwargs)
        self.__name__ = 'swish'
def swish(x,beta=1):
    return (x*sigmoid(x*beta))
get_custom_objects().update({'swish': Swish(swish)})