Timedistribued Flatten after Timedistributed add throws symbolic tensor instance error

conv lstm model.py

def get_model():

    input_image_seqs = Input(shape=(None, 32, 32, 1), name='input_image_seqs')
    bn1 = BatchNormalization()(input_image_seqs)
    ac1 = Activation('elu')(bn1)
    conv1 = TimeDistributed(Conv2D(32, (3, 3), strides=(1, 1), activation='elu', padding='same'))(ac1)
    bn2 = BatchNormalization()(conv1)
    ac2 = Activation('elu')(bn2)
    conv2 = TimeDistributed(Conv2D(32, (3, 3), strides=(1, 1), activation='elu', padding='same'))(ac2)

    residual_shape = K.int_shape(conv2)
    ROW_AXIS = 1
    COL_AXIS = 2
    CHANNEL_AXIS = 3
    shortcut = TimeDistributed(Conv2D(filters=residual_shape[CHANNEL_AXIS],

                          kernel_size=(1, 1),

                          strides=(1, 1),

                          padding="valid",

                          kernel_initializer="he_normal", kernel_regularizer=l2(0.0001)))(input_image_seqs)
    skip1 = TimeDistributed(add([shortcut, conv2]))
    flat1 = TimeDistributed(Flatten())(skip1)
    conv_dense1 = TimeDistributed(Dense(512, activation='elu'))(flat1)

    lstm1 = Bidirectional(LSTM(512, return_sequences=True, activation='elu'), merge_mode='ave')(conv_dense1)
    dense1 = Dense(3, activation=Pitanh)(lstm1)

    model = Model(inputs=[input_image_seqs], outputs=dense1)

    return model
'''
ERROR
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
/usr/local/lib/python3.5/dist-packages/keras/engine/base_layer.py in assert_input_compatibility(self, inputs)
    278             try:
--> 279                 K.is_keras_tensor(x)
    280             except ValueError:

/usr/local/lib/python3.5/dist-packages/keras/backend/tensorflow_backend.py in is_keras_tensor(x)
    470         raise ValueError('Unexpectedly found an instance of type `' +
--> 471                          str(type(x)) + '`. '
    472                          'Expected a symbolic tensor instance.')

ValueError: Unexpectedly found an instance of type `<class 'keras.layers.wrappers.TimeDistributed'>`. Expected a symbolic tensor instance.

During handling of the above exception, another exception occurred:

ValueError                                Traceback (most recent call last)
<ipython-input-85-fe5b9b618138> in <module>()
----> 1 model = get_model()
      2 model.compile(optimizer=Nadam(lr=0.002, beta_1=0.9, beta_2=0.999, 
      3                               epsilon=None, schedule_decay=0.004, clipnorm=0.5),
      4               loss='mse')
      5 model.summary()

<ipython-input-84-1cb8f2a91a9d> in get_model()
     19                           kernel_initializer="he_normal", kernel_regularizer=l2(0.0001)))(input_image_seqs)
     20     skip1 = TimeDistributed(add([shortcut, conv2]))
---> 21     flat1 = TimeDistributed(Flatten())(skip1)
     22     conv_dense1 = TimeDistributed(Dense(512, activation='elu'))(flat1)
     23 #     embed1 = Embedding(input_dim=10, output_dim=100)(flat1)

/usr/local/lib/python3.5/dist-packages/keras/engine/base_layer.py in __call__(self, inputs, **kwargs)
    412                 # Raise exceptions in case the input is not compatible
    413                 # with the input_spec specified in the layer constructor.
--> 414                 self.assert_input_compatibility(inputs)
    415 
    416                 # Collect input shapes to build layer.

/usr/local/lib/python3.5/dist-packages/keras/engine/base_layer.py in assert_input_compatibility(self, inputs)
    283                                  'Received type: ' +
    284                                  str(type(x)) + '. Full input: ' +
--> 285                                  str(inputs) + '. All inputs to the layer '
    286                                  'should be tensors.')
    287 

ValueError: Layer time_distributed_159 was called with an input that isn't a symbolic tensor. Received type: <class 'keras.layers.wrappers.TimeDistributed'>. Full input: [<keras.layers.wrappers.TimeDistributed object at 0x7f9dfe91bdd8>]. All inputs to the layer should be tensors.
'''