layers.py

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import numpy as np
import warnings

from ..engine import Layer, InputSpec
from .. import backend as K
from ..utils import conv_utils
from ..utils.generic_utils import to_list
from .. import regularizers
from .. import constraints
from .. import activations
from .. import initializers


class MaxoutDense(Layer):
    """A dense maxout layer.
    A `MaxoutDense` layer takes the element-wise maximum of
    `nb_feature` `Dense(input_dim, output_dim)` linear layers.
    This allows the layer to learn a convex,
    piecewise linear activation function over the inputs.
    Note that this is a *linear* layer;
    if you wish to apply activation function
    (you shouldn't need to --they are universal function approximators),
    an `Activation` layer must be added after.
    # Arguments
        output_dim: int > 0.
        nb_feature: number of Dense layers to use internally.
        init: name of initialization function for the weights of the layer
            (see [initializations](../initializations.md)),
            or alternatively, Theano function to use for weights
            initialization. This parameter is only relevant
            if you don't pass a `weights` argument.
        weights: list of Numpy arrays to set as initial weights.
            The list should have 2 elements, of shape `(input_dim, output_dim)`
            and (output_dim,) for weights and biases respectively.
        W_regularizer: instance of [WeightRegularizer](../regularizers.md)
            (eg. L1 or L2 regularization), applied to the main weights matrix.
        b_regularizer: instance of [WeightRegularizer](../regularizers.md),
            applied to the bias.
        activity_regularizer: instance of [ActivityRegularizer](../regularizers.md),
            applied to the network output.
        W_constraint: instance of the [constraints](../constraints.md) module
            (eg. maxnorm, nonneg), applied to the main weights matrix.
        b_constraint: instance of the [constraints](../constraints.md) module,
            applied to the bias.
        bias: whether to include a bias
            (i.e. make the layer affine rather than linear).
        input_dim: dimensionality of the input (integer). This argument
            (or alternatively, the keyword argument `input_shape`)
            is required when using this layer as the first layer in a model.
    # Input shape
        2D tensor with shape: `(nb_samples, input_dim)`.
    # Output shape
        2D tensor with shape: `(nb_samples, output_dim)`.
    # References
        - [Maxout Networks](http://arxiv.org/abs/1302.4389)
    """

    def __init__(self, output_dim,
                 nb_feature=4,
                 init='glorot_uniform',
                 weights=None,
                 W_regularizer=None,
                 b_regularizer=None,
                 activity_regularizer=None,
                 W_constraint=None,
                 b_constraint=None,
                 bias=True,
                 input_dim=None,
                 **kwargs):
        warnings.warn('The `MaxoutDense` layer is deprecated '
                      'and will be removed after 06/2017.')
        self.output_dim = output_dim
        self.nb_feature = nb_feature
        self.init = initializers.get(init)

        self.W_regularizer = regularizers.get(W_regularizer)
        self.b_regularizer = regularizers.get(b_regularizer)
        self.activity_regularizer = regularizers.get(activity_regularizer)

        self.W_constraint = constraints.get(W_constraint)
        self.b_constraint = constraints.get(b_constraint)

        self.bias = bias
        self.initial_weights = weights
        self.input_spec = InputSpec(ndim=2)

        self.input_dim = input_dim
        if self.input_dim:
            kwargs['input_shape'] = (self.input_dim,)
        super(MaxoutDense, self).__init__(**kwargs)

    def build(self, input_shape):
        input_dim = input_shape[1]
        self.input_spec = InputSpec(dtype=K.floatx(),
                                    shape=(None, input_dim))

        self.W = self.add_weight(shape=(self.nb_feature, input_dim, self.output_dim),
                                 initializer=self.init,
                                 name='W',
                                 regularizer=self.W_regularizer,
                                 constraint=self.W_constraint)
        if self.bias:
            self.b = self.add_weight(shape=(self.nb_feature, self.output_dim,),
                                     initializer='zero',
                                     name='b',
                                     regularizer=self.b_regularizer,
                                     constraint=self.b_constraint)
        else:
            self.b = None

        if self.initial_weights is not None:
            self.set_weights(self.initial_weights)
            del self.initial_weights
        self.built = True

    def compute_output_shape(self, input_shape):
        assert input_shape and len(input_shape) == 2
        return (input_shape[0], self.output_dim)

    def call(self, x):
        # no activation, this layer is only linear.
        output = K.dot(x, self.W)
        if self.bias:
            output += self.b
        output = K.max(output, axis=1)
        return output

    def get_config(self):
        config = {'output_dim': self.output_dim,
                  'init': initializers.serialize(self.init),
                  'nb_feature': self.nb_feature,
                  'W_regularizer': regularizers.serialize(self.W_regularizer),
                  'b_regularizer': regularizers.serialize(self.b_regularizer),
                  'activity_regularizer':
                      regularizers.serialize(self.activity_regularizer),
                  'W_constraint': constraints.serialize(self.W_constraint),
                  'b_constraint': constraints.serialize(self.b_constraint),
                  'bias': self.bias,
                  'input_dim': self.input_dim}
        base_config = super(MaxoutDense, self).get_config()
        return dict(list(base_config.items()) + list(config.items()))


class Highway(Layer):
    """Densely connected highway network.
    Highway layers are a natural extension of LSTMs to feedforward networks.
    # Arguments
        init: name of initialization function for the weights of the layer
            (see [initializations](../initializations.md)),
            or alternatively, Theano function to use for weights
            initialization. This parameter is only relevant
            if you don't pass a `weights` argument.
        activation: name of activation function to use
            (see [activations](../activations.md)),
            or alternatively, elementwise Theano function.
            If you don't specify anything, no activation is applied
            (ie. "linear" activation: a(x) = x).
        weights: list of Numpy arrays to set as initial weights.
            The list should have 2 elements, of shape `(input_dim, output_dim)`
            and (output_dim,) for weights and biases respectively.
        W_regularizer: instance of [WeightRegularizer](../regularizers.md)
            (eg. L1 or L2 regularization), applied to the main weights matrix.
        b_regularizer: instance of [WeightRegularizer](../regularizers.md),
            applied to the bias.
        activity_regularizer: instance of [ActivityRegularizer](../regularizers.md),
            applied to the network output.
        W_constraint: instance of the [constraints](../constraints.md) module
            (eg. maxnorm, nonneg), applied to the main weights matrix.
        b_constraint: instance of the [constraints](../constraints.md) module,
            applied to the bias.
        bias: whether to include a bias
            (i.e. make the layer affine rather than linear).
        input_dim: dimensionality of the input (integer). This argument
            (or alternatively, the keyword argument `input_shape`)
            is required when using this layer as the first layer in a model.
    # Input shape
        2D tensor with shape: `(nb_samples, input_dim)`.
    # Output shape
        2D tensor with shape: `(nb_samples, input_dim)`.
    # References
        - [Highway Networks](http://arxiv.org/abs/1505.00387v2)
    """

    def __init__(self,
                 init='glorot_uniform',
                 activation=None,
                 weights=None,
                 W_regularizer=None,
                 b_regularizer=None,
                 activity_regularizer=None,
                 W_constraint=None,
                 b_constraint=None,
                 bias=True,
                 input_dim=None,
                 **kwargs):
        warnings.warn('The `Highway` layer is deprecated '
                      'and will be removed after 06/2017.')
        if 'transform_bias' in kwargs:
            kwargs.pop('transform_bias')
            warnings.warn('`transform_bias` argument is deprecated and '
                          'has been removed.')
        self.init = initializers.get(init)
        self.activation = activations.get(activation)

        self.W_regularizer = regularizers.get(W_regularizer)
        self.b_regularizer = regularizers.get(b_regularizer)
        self.activity_regularizer = regularizers.get(activity_regularizer)

        self.W_constraint = constraints.get(W_constraint)
        self.b_constraint = constraints.get(b_constraint)

        self.bias = bias
        self.initial_weights = weights
        self.input_spec = InputSpec(ndim=2)

        self.input_dim = input_dim
        if self.input_dim:
            kwargs['input_shape'] = (self.input_dim,)
        super(Highway, self).__init__(**kwargs)

    def build(self, input_shape):
        input_dim = input_shape[1]
        self.input_spec = InputSpec(dtype=K.floatx(),
                                    shape=(None, input_dim))

        self.W = self.add_weight(shape=(input_dim, input_dim),
                                 initializer=self.init,
                                 name='W',
                                 regularizer=self.W_regularizer,
                                 constraint=self.W_constraint)
        self.W_carry = self.add_weight(shape=(input_dim, input_dim),
                                       initializer=self.init,
                                       name='W_carry')
        if self.bias:
            self.b = self.add_weight(shape=(input_dim,),
                                     initializer='zero',
                                     name='b',
                                     regularizer=self.b_regularizer,
                                     constraint=self.b_constraint)
            self.b_carry = self.add_weight(shape=(input_dim,),
                                           initializer='one',
                                           name='b_carry')
        else:
            self.b_carry = None

        if self.initial_weights is not None:
            self.set_weights(self.initial_weights)
            del self.initial_weights
        self.built = True

    def call(self, x):
        y = K.dot(x, self.W_carry)
        if self.bias:
            y += self.b_carry
        transform_weight = activations.sigmoid(y)
        y = K.dot(x, self.W)
        if self.bias:
            y += self.b
        act = self.activation(y)
        act *= transform_weight
        output = act + (1 - transform_weight) * x
        return output

    def get_config(self):
        config = {'init': initializers.serialize(self.init),
                  'activation': activations.serialize(self.activation),
                  'W_regularizer': regularizers.serialize(self.W_regularizer),
                  'b_regularizer': regularizers.serialize(self.b_regularizer),
                  'activity_regularizer':
                      regularizers.serialize(self.activity_regularizer),
                  'W_constraint': constraints.serialize(self.W_constraint),
                  'b_constraint': constraints.serialize(self.b_constraint),
                  'bias': self.bias,
                  'input_dim': self.input_dim}
        base_config = super(Highway, self).get_config()
        return dict(list(base_config.items()) + list(config.items()))


def AtrousConvolution1D(*args, **kwargs):
    from ..layers import Conv1D
    if 'atrous_rate' in kwargs:
        rate = kwargs.pop('atrous_rate')
    else:
        rate = 1
    kwargs['dilation_rate'] = rate
    warnings.warn('The `AtrousConvolution1D` layer '
                  ' has been deprecated. Use instead '
                  'the `Conv1D` layer with the `dilation_rate` '
                  'argument.')
    return Conv1D(*args, **kwargs)


def AtrousConvolution2D(*args, **kwargs):
    from ..layers import Conv2D
    if 'atrous_rate' in kwargs:
        rate = kwargs.pop('atrous_rate')
    else:
        rate = 1
    kwargs['dilation_rate'] = rate
    warnings.warn('The `AtrousConvolution2D` layer '
                  ' has been deprecated. Use instead '
                  'the `Conv2D` layer with the `dilation_rate` '
                  'argument.')
    return Conv2D(*args, **kwargs)


class Recurrent(Layer):
    """Abstract base class for recurrent layers.

    Do not use in a model -- it's not a valid layer!
    Use its children classes `LSTM`, `GRU` and `SimpleRNN` instead.
    All recurrent layers (`LSTM`, `GRU`, `SimpleRNN`) also
    follow the specifications of this class and accept
    the keyword arguments listed below.

    # Example

    ```python
        # as the first layer in a Sequential model
        model = Sequential()
        model.add(LSTM(32, input_shape=(10, 64)))
        # now model.output_shape == (None, 32)
        # note: `None` is the batch dimension.
        # for subsequent layers, no need to specify the input size:
        model.add(LSTM(16))
        # to stack recurrent layers, you must use return_sequences=True
        # on any recurrent layer that feeds into another recurrent layer.
        # note that you only need to specify the input size on the first layer.
        model = Sequential()
        model.add(LSTM(64, input_dim=64, input_length=10, return_sequences=True))
        model.add(LSTM(32, return_sequences=True))
        model.add(LSTM(10))
    ```

    # Arguments
        weights: list of Numpy arrays to set as initial weights.
            The list should have 3 elements, of shapes:
            `[(input_dim, output_dim), (output_dim, output_dim), (output_dim,)]`.
        return_sequences: Boolean. Whether to return the last output
            in the output sequence, or the full sequence.
        return_state: Boolean. Whether to return the last state
            in addition to the output.
        go_backwards: Boolean (default False).
            If True, process the input sequence backwards and return the
            reversed sequence.
        stateful: Boolean (default False). If True, the last state
            for each sample at index i in a batch will be used as initial
            state for the sample of index i in the following batch.
        unroll: Boolean (default False).
            If True, the network will be unrolled,
            else a symbolic loop will be used.
            Unrolling can speed-up a RNN,
            although it tends to be more memory-intensive.
            Unrolling is only suitable for short sequences.
        implementation: one of {0, 1, or 2}.
            If set to 0, the RNN will use
            an implementation that uses fewer, larger matrix products,
            thus running faster on CPU but consuming more memory.
            If set to 1, the RNN will use more matrix products,
            but smaller ones, thus running slower
            (may actually be faster on GPU) while consuming less memory.
            If set to 2 (LSTM/GRU only),
            the RNN will combine the input gate,
            the forget gate and the output gate into a single matrix,
            enabling more time-efficient parallelization on the GPU.
            Note: RNN dropout must be shared for all gates,
            resulting in a slightly reduced regularization.
        input_dim: dimensionality of the input (integer).
            This argument (or alternatively, the keyword argument `input_shape`)
            is required when using this layer as the first layer in a model.
        input_length: Length of input sequences, to be specified
            when it is constant.
            This argument is required if you are going to connect
            `Flatten` then `Dense` layers upstream
            (without it, the shape of the dense outputs cannot be computed).
            Note that if the recurrent layer is not the first layer
            in your model, you would need to specify the input length
            at the level of the first layer
            (e.g. via the `input_shape` argument)

    # Input shapes
        3D tensor with shape `(batch_size, timesteps, input_dim)`,
        (Optional) 2D tensors with shape `(batch_size, output_dim)`.

    # Output shape
        - if `return_state`: a list of tensors. The first tensor is
            the output. The remaining tensors are the last states,
            each with shape `(batch_size, units)`.
        - if `return_sequences`: 3D tensor with shape
            `(batch_size, timesteps, units)`.
        - else, 2D tensor with shape `(batch_size, units)`.

    # Masking
        This layer supports masking for input data with a variable number
        of timesteps. To introduce masks to your data,
        use an [Embedding](embeddings.md) layer with the `mask_zero` parameter
        set to `True`.

    # Note on using statefulness in RNNs
        You can set RNN layers to be 'stateful', which means that the states
        computed for the samples in one batch will be reused as initial states
        for the samples in the next batch. This assumes a one-to-one mapping
        between samples in different successive batches.
        To enable statefulness:
            - specify `stateful=True` in the layer constructor.
            - specify a fixed batch size for your model, by passing
                if sequential model:
                  `batch_input_shape=(...)` to the first layer in your model.
                else for functional model with 1 or more Input layers:
                  `batch_shape=(...)` to all the first layers in your model.
                This is the expected shape of your inputs
                *including the batch size*.
                It should be a tuple of integers, e.g. `(32, 10, 100)`.
            - specify `shuffle=False` when calling fit().
        To reset the states of your model, call `.reset_states()` on either
        a specific layer, or on your entire model.

    # Note on specifying the initial state of RNNs
        You can specify the initial state of RNN layers symbolically by
        calling them with the keyword argument `initial_state`. The value of
        `initial_state` should be a tensor or list of tensors representing
        the initial state of the RNN layer.
        You can specify the initial state of RNN layers numerically by
        calling `reset_states` with the keyword argument `states`. The value of
        `states` should be a numpy array or list of numpy arrays representing
        the initial state of the RNN layer.
    """

    def __init__(self, return_sequences=False,
                 return_state=False,
                 go_backwards=False,
                 stateful=False,
                 unroll=False,
                 implementation=0,
                 **kwargs):
        super(Recurrent, self).__init__(**kwargs)
        self.return_sequences = return_sequences
        self.return_state = return_state
        self.go_backwards = go_backwards

        self.stateful = stateful
        self.unroll = unroll
        self.implementation = implementation
        self.supports_masking = True
        self.input_spec = [InputSpec(ndim=3)]
        self.state_spec = None
        self.dropout = 0
        self.recurrent_dropout = 0

    def compute_output_shape(self, input_shape):
        if isinstance(input_shape, list):
            input_shape = input_shape[0]

        if self.return_sequences:
            output_shape = (input_shape[0], input_shape[1], self.units)
        else:
            output_shape = (input_shape[0], self.units)

        if self.return_state:
            state_shape = [(input_shape[0], self.units) for _ in self.states]
            return [output_shape] + state_shape
        else:
            return output_shape

    def compute_mask(self, inputs, mask):
        if isinstance(mask, list):
            mask = mask[0]
        output_mask = mask if self.return_sequences else None
        if self.return_state:
            state_mask = [None for _ in self.states]
            return [output_mask] + state_mask
        else:
            return output_mask

    def step(self, inputs, states):
        raise NotImplementedError

    def get_constants(self, inputs, training=None):
        return []

    def get_initial_state(self, inputs):
        # build an all-zero tensor of shape (samples, output_dim)
        initial_state = K.zeros_like(inputs)  # (samples, timesteps, input_dim)
        initial_state = K.sum(initial_state, axis=(1, 2))  # (samples,)
        initial_state = K.expand_dims(initial_state)  # (samples, 1)
        # (samples, output_dim)
        initial_state = K.tile(initial_state, [1, self.units])
        initial_state = [initial_state for _ in range(len(self.states))]
        return initial_state

    def preprocess_input(self, inputs, training=None):
        return inputs

    def __call__(self, inputs, initial_state=None, **kwargs):

        # If there are multiple inputs, then
        # they should be the main input and `initial_state`
        # e.g. when loading model from file
        if (isinstance(inputs, (list, tuple))
                and len(inputs) > 1 and initial_state is None):
            initial_state = inputs[1:]
            inputs = inputs[0]

        # If `initial_state` is specified,
        # and if it a Keras tensor,
        # then add it to the inputs and temporarily
        # modify the input spec to include the state.
        if initial_state is None:
            return super(Recurrent, self).__call__(inputs, **kwargs)

        initial_state = to_list(initial_state, allow_tuple=True)

        is_keras_tensor = hasattr(initial_state[0], '_keras_history')
        for tensor in initial_state:
            if hasattr(tensor, '_keras_history') != is_keras_tensor:
                raise ValueError('The initial state of an RNN layer cannot be'
                                 ' specified with a mix of Keras tensors and'
                                 ' non-Keras tensors')

        if is_keras_tensor:
            # Compute the full input spec, including state
            input_spec = self.input_spec
            state_spec = self.state_spec
            input_spec = to_list(input_spec)
            state_spec = to_list(state_spec)
            self.input_spec = input_spec + state_spec

            # Compute the full inputs, including state
            inputs = [inputs] + list(initial_state)

            # Perform the call
            output = super(Recurrent, self).__call__(inputs, **kwargs)

            # Restore original input spec
            self.input_spec = input_spec
            return output
        else:
            kwargs['initial_state'] = initial_state
            return super(Recurrent, self).__call__(inputs, **kwargs)

    def call(self, inputs, mask=None, training=None, initial_state=None):
        # input shape: `(samples, time (padded with zeros), input_dim)`
        # note that the .build() method of subclasses MUST define
        # self.input_spec and self.state_spec with complete input shapes.
        if isinstance(inputs, list):
            initial_state = inputs[1:]
            inputs = inputs[0]
        elif initial_state is not None:
            pass
        elif self.stateful:
            initial_state = self.states
        else:
            initial_state = self.get_initial_state(inputs)

        if isinstance(mask, list):
            mask = mask[0]

        if len(initial_state) != len(self.states):
            raise ValueError('Layer has ' + str(len(self.states)) +
                             ' states but was passed ' +
                             str(len(initial_state)) +
                             ' initial states.')
        input_shape = K.int_shape(inputs)
        timesteps = input_shape[1]
        if self.unroll and timesteps in [None, 1]:
            raise ValueError('Cannot unroll a RNN if the '
                             'time dimension is undefined or equal to 1. \n'
                             '- If using a Sequential model, '
                             'specify the time dimension by passing '
                             'an `input_shape` or `batch_input_shape` '
                             'argument to your first layer. If your '
                             'first layer is an Embedding, you can '
                             'also use the `input_length` argument.\n'
                             '- If using the functional API, specify '
                             'the time dimension by passing a `shape` '
                             'or `batch_shape` argument to your Input layer.')
        constants = self.get_constants(inputs, training=None)
        preprocessed_input = self.preprocess_input(inputs, training=None)
        last_output, outputs, states = K.rnn(self.step,
                                             preprocessed_input,
                                             initial_state,
                                             go_backwards=self.go_backwards,
                                             mask=mask,
                                             constants=constants,
                                             unroll=self.unroll,
                                             input_length=timesteps)
        if self.stateful:
            updates = []
            for i in range(len(states)):
                updates.append((self.states[i], states[i]))
            self.add_update(updates, inputs)

        # Properly set learning phase
        if 0 < self.dropout + self.recurrent_dropout:
            last_output._uses_learning_phase = True
            outputs._uses_learning_phase = True

        if self.return_sequences:
            output = outputs
        else:
            output = last_output

        if self.return_state:
            states = to_list(states, allow_tuple=True)
            return [output] + states
        else:
            return output

    def reset_states(self, states=None):
        if not self.stateful:
            raise AttributeError('Layer must be stateful.')
        batch_size = self.input_spec[0].shape[0]
        if not batch_size:
            raise ValueError('If a RNN is stateful, it needs to know '
                             'its batch size. Specify the batch size '
                             'of your input tensors: \n'
                             '- If using a Sequential model, '
                             'specify the batch size by passing '
                             'a `batch_input_shape` '
                             'argument to your first layer.\n'
                             '- If using the functional API, specify '
                             'the time dimension by passing a '
                             '`batch_shape` argument to your Input layer.')
        # initialize state if None
        if self.states[0] is None:
            self.states = [K.zeros((batch_size, self.units))
                           for _ in self.states]
        elif states is None:
            for state in self.states:
                K.set_value(state, np.zeros((batch_size, self.units)))
        else:
            states = to_list(states, allow_tuple=True)
            if len(states) != len(self.states):
                raise ValueError('Layer ' + self.name + ' expects ' +
                                 str(len(self.states)) + ' states, '
                                 'but it received ' + str(len(states)) +
                                 ' state values. Input received: ' +
                                 str(states))
            for index, (value, state) in enumerate(zip(states, self.states)):
                if value.shape != (batch_size, self.units):
                    raise ValueError('State ' + str(index) +
                                     ' is incompatible with layer ' +
                                     self.name + ': expected shape=' +
                                     str((batch_size, self.units)) +
                                     ', found shape=' + str(value.shape))
                K.set_value(state, value)

    def get_config(self):
        config = {'return_sequences': self.return_sequences,
                  'return_state': self.return_state,
                  'go_backwards': self.go_backwards,
                  'stateful': self.stateful,
                  'unroll': self.unroll,
                  'implementation': self.implementation}
        base_config = super(Recurrent, self).get_config()
        return dict(list(base_config.items()) + list(config.items()))


class ConvRecurrent2D(Recurrent):
    """Abstract base class for convolutional recurrent layers.

    Do not use in a model -- it's not a functional layer!

    # Arguments
        filters: Integer, the dimensionality of the output space
            (i.e. the number output of filters in the convolution).
        kernel_size: An integer or tuple/list of n integers, specifying the
            dimensions of the convolution window.
        strides: An integer or tuple/list of n integers,
            specifying the strides of the convolution.
            Specifying any stride value != 1 is incompatible with specifying
            any `dilation_rate` value != 1.
        padding: One of `"valid"` or `"same"` (case-insensitive).
        data_format: A string,
            one of `channels_last` (default) or `channels_first`.
            The ordering of the dimensions in the inputs.
            `channels_last` corresponds to inputs with shape
            `(batch, time, ..., channels)`
            while `channels_first` corresponds to
            inputs with shape `(batch, time, channels, ...)`.
            It defaults to the `image_data_format` value found in your
            Keras config file at `~/.keras/keras.json`.
            If you never set it, then it will be "channels_last".
        dilation_rate: An integer or tuple/list of n integers, specifying
            the dilation rate to use for dilated convolution.
            Currently, specifying any `dilation_rate` value != 1 is
            incompatible with specifying any `strides` value != 1.
        return_sequences: Boolean. Whether to return the last output
            in the output sequence, or the full sequence.
        go_backwards: Boolean (default False).
            If True, process the input sequence backwards.
        stateful: Boolean (default False). If True, the last state
            for each sample at index i in a batch will be used as initial
            state for the sample of index i in the following batch.

    # Input shape
        5D tensor with shape `(num_samples, timesteps, channels, rows, cols)`.

    # Output shape
        - if `return_sequences`: 5D tensor with shape
            `(num_samples, timesteps, channels, rows, cols)`.
        - else, 4D tensor with shape `(num_samples, channels, rows, cols)`.

    # Masking
        This layer supports masking for input data with a variable number
        of timesteps. To introduce masks to your data,
        use an [Embedding](embeddings.md) layer with the `mask_zero` parameter
        set to `True`.
        **Note:** for the time being, masking is only supported with Theano.

    # Note on using statefulness in RNNs
        You can set RNN layers to be 'stateful', which means that the states
        computed for the samples in one batch will be reused as initial states
        for the samples in the next batch.
        This assumes a one-to-one mapping between
        samples in different successive batches.

        To enable statefulness:
            - specify `stateful=True` in the layer constructor.
            - specify a fixed batch size for your model, by passing
                a `batch_input_size=(...)` to the first layer in your model.
                This is the expected shape of your inputs *including the batch
                size*.
                It should be a tuple of integers, e.g. `(32, 10, 100)`.

        To reset the states of your model, call `.reset_states()` on either
        a specific layer, or on your entire model.
    """

    def __init__(self, filters,
                 kernel_size,
                 strides=(1, 1),
                 padding='valid',
                 data_format=None,
                 dilation_rate=(1, 1),
                 return_sequences=False,
                 go_backwards=False,
                 stateful=False,
                 **kwargs):
        super(ConvRecurrent2D, self).__init__(**kwargs)
        self.filters = filters
        self.kernel_size = conv_utils.normalize_tuple(kernel_size, 2, 'kernel_size')
        self.strides = conv_utils.normalize_tuple(strides, 2, 'strides')
        self.padding = conv_utils.normalize_padding(padding)
        self.data_format = K.normalize_data_format(data_format)
        self.dilation_rate = conv_utils.normalize_tuple(dilation_rate, 2,
                                                        'dilation_rate')
        self.return_sequences = return_sequences
        self.go_backwards = go_backwards
        self.stateful = stateful
        self.input_spec = [InputSpec(ndim=5)]
        self.state_spec = None

    def compute_output_shape(self, input_shape):
        if isinstance(input_shape, list):
            input_shape = input_shape[0]
        if self.data_format == 'channels_first':
            rows = input_shape[3]
            cols = input_shape[4]
        elif self.data_format == 'channels_last':
            rows = input_shape[2]
            cols = input_shape[3]
        rows = conv_utils.conv_output_length(rows,
                                             self.kernel_size[0],
                                             padding=self.padding,
                                             stride=self.strides[0],
                                             dilation=self.dilation_rate[0])
        cols = conv_utils.conv_output_length(cols,
                                             self.kernel_size[1],
                                             padding=self.padding,
                                             stride=self.strides[1],
                                             dilation=self.dilation_rate[1])
        if self.return_sequences:
            if self.data_format == 'channels_first':
                output_shape = (input_shape[0], input_shape[1],
                                self.filters, rows, cols)
            elif self.data_format == 'channels_last':
                output_shape = (input_shape[0], input_shape[1],
                                rows, cols, self.filters)
        else:
            if self.data_format == 'channels_first':
                output_shape = (input_shape[0], self.filters, rows, cols)
            elif self.data_format == 'channels_last':
                output_shape = (input_shape[0], rows, cols, self.filters)

        if self.return_state:
            if self.data_format == 'channels_first':
                state_shape = (input_shape[0], self.filters, rows, cols)
            elif self.data_format == 'channels_last':
                state_shape = (input_shape[0], rows, cols, self.filters)
            output_shape = [output_shape, state_shape, state_shape]

        return output_shape

    def get_config(self):
        config = {'filters': self.filters,
                  'kernel_size': self.kernel_size,
                  'strides': self.strides,
                  'padding': self.padding,
                  'data_format': self.data_format,
                  'dilation_rate': self.dilation_rate,
                  'return_sequences': self.return_sequences,
                  'go_backwards': self.go_backwards,
                  'stateful': self.stateful}
        base_config = super(ConvRecurrent2D, self).get_config()
        return dict(list(base_config.items()) + list(config.items()))