keras-team/keras · error · ValueError

All cells must have a `call` method. Received cell without a

Error message

All cells must have a `call` method. Received cell without a `call` method: {cell}

What it means

StackedRNNCells wraps a list of RNN cell objects (e.g. SimpleRNNCell instances) that an RNN layer steps through. Each cell must be a duck-typed RNN cell exposing both a `call` method and a `state_size` attribute; `__init__` checks every element with `dir(cell)` and rejects any object missing `call`. This fails at construction, before any computation.

Source

Thrown at keras/src/layers/rnn/stacked_rnn_cells.py:38

    batch_size = 3
    sentence_length = 5
    num_features = 2
    new_shape = (batch_size, sentence_length, num_features)
    x = np.reshape(np.arange(30), new_shape)

    rnn_cells = [keras.layers.LSTMCell(128) for _ in range(2)]
    stacked_lstm = keras.layers.StackedRNNCells(rnn_cells)
    lstm_layer = keras.layers.RNN(stacked_lstm)

    result = lstm_layer(x)
    ```
    """

    def __init__(self, cells, **kwargs):
        super().__init__(**kwargs)
        for cell in cells:
            if "call" not in dir(cell):
                raise ValueError(
                    "All cells must have a `call` method. "
                    f"Received cell without a `call` method: {cell}"
                )
            if "state_size" not in dir(cell):
                raise ValueError(
                    "All cells must have a `state_size` attribute. "
                    f"Received cell without a `state_size`: {cell}"
                )
        self.cells = cells

    @property
    def state_size(self):
        return [c.state_size for c in self.cells]

    @property
    def output_size(self):
        if getattr(self.cells[-1], "output_size", None) is not None:
            return self.cells[-1].output_size

View on GitHub (pinned to 7a34a03db6)

Solutions

1. Pass instantiated cell objects: StackedRNNCells(cells=[SimpleRNNCell(8), GRUCell(8)])
2. For custom cells, implement both call(self, inputs, states) returning (output, new_states) and a state_size attribute
3. Do not pass Layers or dicts; convert them to proper Keras cells first

Example fix

# before
cells = keras.layers.StackedRNNCells([keras.layers.SimpleRNNCell, keras.layers.GRUCell])

# after
cells = keras.layers.StackedRNNCells([keras.layers.SimpleRNNCell(8), keras.layers.GRUCell(8)])

assert all(hasattr(c, 'call') and hasattr(c, 'state_size') for c in cells), 'all entries must be RNN cells'
stacked = keras.layers.StackedRNNCells(cells=cells)

def is_rnn_cell(obj) -> bool:
    return hasattr(obj, 'call') and hasattr(obj, 'state_size')

When it happens

Trigger: Passing plain objects, dicts, Keras Layer instances that are not cells, or class objects (instead of instantiated cell objects) to keras.layers.StackedRNNCells(cells=[...]). E.g. cells=[SimpleRNNCell] (class, not instance) or cells=[{'units': 8}].

Common situations: Migrating tf.keras code where cell lists were built differently; forgetting parentheses when instantiating cells; passing a Layer subclass that lacks the RNN cell interface; wrapping non-cell custom layers in a stacked RNN.

Related errors

• Received an invalid value for argument `units`, expected a p
• All cells must have a `state_size` attribute. Received cell
• Unknown activation function '{activation}' cannot be seriali
• Could not interpret activation function identifier: {identif
• ConvNeXt does not support the `channels_first` image data fo

AI-assisted analysis of keras-team/keras@7a34a03db6 (2026-08-25). Data as JSON: /api/errors/757a5c16dd32c7ab. Report an issue: GitHub.