我的顺序模型的性能比Keras的功能性差

Question

昨天，我终于可以在Keras实施我的模型了。但是对于我的学习，我想在Sequential和Functional中实现我的模型。 Convert Sequential to Functional in Keras

我的Functional实际上运作良好。另一方面，Sequential有效，但准确性比Functional差。

你能在我的Sequential模型中发现任何错误吗？

if is_sequential:
    # Sequential, working
    x = Sequential()
    x.add(LSTM(128, 
               input_shape=(seq_len, input_dim),
               activation='tanh', 
               return_sequences=True))
    x.add(TimeDistributed(Dense(out_size, activation='softmax')))
    mask= Sequential()
    mask.add(Dense(out_size, input_shape=(seq_len, out_size)))
    model = Sequential()
    model.add(Merge([x, mask], mode='mul'))
else:
    # Functional, working
    x = Input((seq_len, input_dim))
    lstm = LSTM(128, return_sequences=True, activation='tanh')(x)
    td = TimeDistributed(Dense(out_size, activation='softmax'))(lstm)
    mask= Input((seq_len, out_size))
    out = merge([td, mask], mode='mul')
    model = Model(input=[x, mask], output=out)

# same in Sequential and Functional
model.compile(...) 
model.fit(...)