Non-Stateful LSTM Issues with Keras - keras

Question

Good day,
I am trying to create a LSTM model (stateful or non-stateful) but running into several issues.
I am attempting to add a layer using:
model = Sequential()
...
model.add(LSTM(c['num_rnn_unit'],
activation=c['rnn_activation'],
dropout=c['dropout_rnn_input'],
recurrent_dropout=c['dropout_rnn_recurrent'],
return_sequences=True,
stateful=False,
#batch_input_shape=(c['batch_size'],c['num_steps'], c['input_dim'])
))
where:
'num_rnn_unit': np.random.choice([16, 32, 64, 128, 256, 512, 1024])
'rnn_activation': np.random.choice(['tanh', 'sigmoid'])
'dropout_rnn_input': 0
'batch_size': np.random.choice([64, 128])
'num_steps':np.random.choice([5, 10, 15])
'input_dim': 64
I experimented with "stateful=True" and used the commented out "batch_input_shape" but this caused additional errors, which others have had as well but found no workable solution.
So I stuck with trying to make "stateful=False" to work but it yields the error (below).
Any thoughts on why this error is coming up? Thanks in advance!
Here is the traceback:
~\AppData\Local\Continuum\anaconda3\envs\env_py3\lib\site-packages\keras\engine\sequential.py in add(self, layer)
180 self.inputs = network.get_source_inputs(self.outputs[0])
181 elif self.outputs:
--> 182 output_tensor = layer(self.outputs[0])
183 if isinstance(output_tensor, list):
184 raise TypeError('All layers in a Sequential model '
~\AppData\Local\Continuum\anaconda3\envs\env_py3\lib\site-packages\keras\layers\recurrent.py in __call__(self, inputs, initial_state, constants, **kwargs)
539
540 if initial_state is None and constants is None:
--> 541 return super(RNN, self).__call__(inputs, **kwargs)
542
543 # If any of `initial_state` or `constants` are specified and are Keras
~\AppData\Local\Continuum\anaconda3\envs\env_py3\lib\site-packages\keras\backend\tensorflow_backend.py in symbolic_fn_wrapper(*args, **kwargs)
73 if _SYMBOLIC_SCOPE.value:
74 with get_graph().as_default():
---> 75 return func(*args, **kwargs)
76 else:
77 return func(*args, **kwargs)
~\AppData\Local\Continuum\anaconda3\envs\env_py3\lib\site-packages\keras\engine\base_layer.py in __call__(self, inputs, **kwargs)
487 # Actually call the layer,
488 # collecting output(s), mask(s), and shape(s).
--> 489 output = self.call(inputs, **kwargs)
490 output_mask = self.compute_mask(inputs, previous_mask)
491
~\AppData\Local\Continuum\anaconda3\envs\env_py3\lib\site-packages\keras\layers\recurrent.py in call(self, inputs, mask, training, initial_state)
1689 mask=mask,
1690 training=training,
-> 1691 initial_state=initial_state)
1692
1693 #property
~\AppData\Local\Continuum\anaconda3\envs\env_py3\lib\site-packages\keras\layers\recurrent.py in call(self, inputs, mask, training, initial_state, constants)
635 mask = mask[0]
636
--> 637 if len(initial_state) != len(self.states):
638 raise ValueError('Layer has ' + str(len(self.states)) +
639 ' states but was passed ' +
~\AppData\Local\Continuum\anaconda3\envs\env_py3\lib\site-packages\keras\layers\recurrent.py in states(self)
436 num_states = 1
437 else:
--> 438 num_states = len(self.cell.state_size)
439 return [None for _ in range(num_states)]
440 return self._states
TypeError: object of type 'numpy.int32' has no len()
Would this be the first layer "input_shape", with batch_normalization=True:
if c['batch_normalization']:
model.add(BatchNormalization(input_shape=(c['num_steps'], c['input_dim'])))
model.add(TimeDistributed(Dropout(c['dropout_input']),
input_shape=(c['num_steps'], c['input_dim'])))