TypeError：super（）在python 2和python 3中至少接受1个参数（给定0）

Question

我在超级方法参数中添加了子类和self，仍然         错误仍然存​​在。您可以提出任何建议吗？此外，我尝试在python 2和python 3上都运行代码，但是仍然存在相同的问题。

TypeError                                 Traceback (most recent call last)
<ipython-input-66-4a9264e84ddd> in <module>()
49 model = Sequential()
50 model.add(LSTM(150, input_shape=(n_timesteps_in, n_features), 
return_sequences=True))
---> 51 model.add(AttentionDecoder(127, n_features))
52 model.compile(loss='categorical_crossentropy', optimizer='adam', metrics= 
['acc'])
53 # train LSTM

 51 
 52         super(AttentionDecoder, self). __init__(units, output_dim)
 ---> 53         self.name = name
 54         self.return_sequences = True  # must return sequences
 55 

TypeError: super() takes at least 1 argument (0 given)





class AttentionDecoder(Recurrent):

def __init__(self, units, output_dim,
             activation='tanh',
             return_probabilities=False,
             name='AttentionDecoder',
             kernel_initializer='glorot_uniform',
             recurrent_initializer='orthogonal',
             bias_initializer='zeros',
             kernel_regularizer=None,
             bias_regularizer=None,
             activity_regularizer=None,
             kernel_constraint=None,
             bias_constraint=None,
             **kwargs):


    super(AttentionDecoder, self). __init__(units, output_dim)
    self.name = name
    self.return_sequences = True  # must return sequences