Python tensorflow.python.ops.rnn_cell.LSTMStateTuple() Examples

def __call__(self, inputs, state, k_size=3, scope=None):
    """Convolutional Long short-term memory cell (ConvLSTM)."""
    with vs.variable_scope(scope or type(self).__name__): # "ConvLSTMCell"
      if self._state_is_tuple:
        c, h = state
      else:
        c, h = array_ops.split(3, 2, state)

      # batch_size * height * width * channel
      concat = _conv([inputs, h], 4 * self._num_units, k_size, True)

      # i = input_gate, j = new_input, f = forget_gate, o = output_gate
      i, j, f, o = array_ops.split(3, 4, concat)

      new_c = (c * sigmoid(f + self._forget_bias) + sigmoid(i) *
               self._activation(j))
      new_h = self._activation(new_c) * sigmoid(o)

      if self._state_is_tuple:
        new_state = LSTMStateTuple(new_c, new_h)
      else:
        new_state = array_ops.concat(3, [new_c, new_h])
      return new_h, new_state 

def __call__(self, inputs, state, scope=None):
    """Convolutional Long short-term memory cell (ConvLSTM)."""
    with vs.variable_scope(scope or type(self).__name__): # "ConvLSTMCell"
      if self._state_is_tuple:
        c, h = state
      else:
        c, h = array_ops.split(3, 2, state)

      # batch_size * height * width * channel
      concat = _conv([inputs, h], 4 * self._num_units, self._k_size, True, initializer=self._initializer)

      # i = input_gate, j = new_input, f = forget_gate, o = output_gate
      i, j, f, o = array_ops.split(3, 4, concat)

      new_c = (c * sigmoid(f + self._forget_bias) + sigmoid(i) *
               self._activation(j))
      new_h = self._activation(new_c) * sigmoid(o)

      if self._state_is_tuple:
        new_state = LSTMStateTuple(new_c, new_h)
      else:
        new_state = array_ops.concat(3, [new_c, new_h])
      return new_h, new_state 

def __call__(self, inputs, state, scope=None):
    """Run the cell with the declared zoneouts."""

    # compute output and new state as before
    output, new_state = self._cell(inputs, state, scope)

    # if either hidden state or memory cell zoneout is applied, then split state and process
    if self._has_hidden_state_zoneout or self._has_memory_cell_zoneout:
      # split state
      c_old, m_old = state
      c_new, m_new = new_state

      # apply zoneout to memory cell and hidden state
      c_and_m = []
      for s_old, s_new, p, has_zoneout in [(c_old, c_new, self._memory_cell_keep_prob,  self._has_memory_cell_zoneout), 
                                           (m_old, m_new, self._hidden_state_keep_prob, self._has_hidden_state_zoneout)]:
        if has_zoneout:
          if self._is_training:
            mask = nn_ops.dropout(array_ops.ones_like(s_new), p, seed=self._seed) * p # this should just random ops instead. See dropout code for how.
            s = ((1. - mask) * s_old) + (mask * s_new)
          else:
            s = ((1. - p) * s_old) + (p * s_new)
        else:
          s = s_new

        c_and_m.append(s)

      # package final results
      new_state = LSTMStateTuple(*c_and_m)
      output = new_state.h

    return output, new_state 

def state_size(self):
    return (LSTMStateTuple(self._num_units, self._num_units)
            if self._state_is_tuple else 2 * self._num_units) 

def state_size(self):
    return rnn_cell.LSTMStateTuple(self._num_units, self._num_units) 

def __call__(self, inputs, state, scope=None):
    """LSTM cell with layer normalization and recurrent dropout."""

    with vs.variable_scope(scope or type(self).__name__) as scope:  # LayerNormBasicLSTMCell  # pylint: disable=unused-variables
      c, h = state
      args = array_ops.concat(1, [inputs, h])
      concat = self._linear(args)

      i, j, f, o = array_ops.split(1, 4, concat)
      if self._layer_norm:
        i = self._norm(i, "input")
        j = self._norm(j, "transform")
        f = self._norm(f, "forget")
        o = self._norm(o, "output")

      g = self._activation(j)
      if (not isinstance(self._keep_prob, float)) or self._keep_prob < 1:
        g = nn_ops.dropout(g, self._keep_prob, seed=self._seed)

      new_c = (c * math_ops.sigmoid(f + self._forget_bias)
               + math_ops.sigmoid(i) * g)
      if self._layer_norm:
        new_c = self._norm(new_c, "state")
      new_h = self._activation(new_c) * math_ops.sigmoid(o)

      new_state = rnn_cell.LSTMStateTuple(new_c, new_h)
      return new_h, new_state 

def state_size(self):
    return (LSTMStateTuple(self._num_units, self._num_units)
            if self._state_is_tuple else 2 * self._num_units) 

def state_size(self):
    return (LSTMStateTuple(self._num_units, self._num_units)
            if self._state_is_tuple else 2 * self._num_units) 

def __call__(self, inputs, state, scope=None):
    """Convolutional Long short-term memory cell (ConvLSTM)."""
    with vs.variable_scope(scope or type(self).__name__): # "ConvLSTMCell"
      if self._state_is_tuple:
        c, h = state
      else:
        c, h = array_ops.split(3, 2, state)
      s1 = vs.get_variable("s1", initializer=tf.ones([self._height, self._width, 4 * self._num_units]), dtype=tf.float32)
      s2 = vs.get_variable("s2", initializer=tf.ones([self._height, self._width, 4 * self._num_units]), dtype=tf.float32)
      # s3 = vs.get_variable("s3", initializer=tf.ones([self._batch_size, self._num_units]), dtype=tf.float32)

      b1 = vs.get_variable("b1", initializer=tf.zeros([self._height, self._width, 4 * self._num_units]), dtype=tf.float32)
      b2 = vs.get_variable("b2", initializer=tf.zeros([self._height, self._width, 4 * self._num_units]), dtype=tf.float32)
      # b3 = vs.get_variable("b3", initializer=tf.zeros([self._batch_size, self._num_units]), dtype=tf.float32)
      input_below_ = _conv([inputs], 4 * self._num_units, self._k_size, False, initializer=self._initializer, scope="out_1")
      input_below_ = ln(input_below_, s1, b1)
      state_below_ = _conv([h], 4 * self._num_units, self._k_size, False, initializer=self._initializer, scope="out_2")
      state_below_ = ln(state_below_, s2, b2)
      lstm_matrix = tf.add(input_below_, state_below_)

      i, j, f, o = array_ops.split(3, 4, lstm_matrix)

      # batch_size * height * width * channel
      # concat = _conv([inputs, h], 4 * self._num_units, self._k_size, True, initializer=self._initializer)

      # i = input_gate, j = new_input, f = forget_gate, o = output_gate
      # i, j, f, o = array_ops.split(3, 4, lstm_matrix)

      new_c = (c * sigmoid(f + self._forget_bias) + sigmoid(i) *
               self._activation(j))
      new_h = self._activation(new_c) * sigmoid(o)

      if self._state_is_tuple:
        new_state = LSTMStateTuple(new_c, new_h)
      else:
        new_state = array_ops.concat(3, [new_c, new_h])
      return new_h, new_state 

def state_size(self):
    return (rnn_cell.LSTMStateTuple(self._num_units, self._num_units)
            if self._state_is_tuple else 2 * self._num_units) 

def call(self, inputs, state):
    """Long short-term memory cell (LSTM).

    Args:
      inputs: `2-D` tensor with shape `[batch_size x input_size]`.
      state: An `LSTMStateTuple` of state tensors, each shaped
        `[batch_size x self.state_size]`, if `state_is_tuple` has been set to
        `True`.  Otherwise, a `Tensor` shaped
        `[batch_size x 2 * self.state_size]`.

    Returns:
      A pair containing the new hidden state, and the new state (either a
        `LSTMStateTuple` or a concatenated state, depending on
        `state_is_tuple`).
    """
    sigmoid = math_ops.sigmoid
    # Parameters of gates are concatenated into one multiply for efficiency.
    if self._state_is_tuple:
      c, h = state
    else:
      c, h = array_ops.split(value=state, num_or_size_splits=2, axis=1)

    # concat = _linear([inputs, h], 4 * self._num_units, True)
    concat = _linear(inputs, 4 * self._num_units, True)

    # i = input_gate, j = new_input, f = forget_gate, o = output_gate
    i, j, f, o = array_ops.split(value=concat, num_or_size_splits=4, axis=1)

    new_c = (
        c * sigmoid(f + self._forget_bias) + sigmoid(i) * self._activation(j))
    new_h = self._activation(new_c) * sigmoid(o)

    if self._state_is_tuple:
      new_state = rnn_cell.LSTMStateTuple(new_c, new_h)
    else:
      new_state = array_ops.concat([new_c, new_h], 1)
    return new_h, new_state 

def __init__(self, num_units, use_peepholes=False,
               initializer=None, num_proj=None, proj_clip=None,
               num_unit_shards=1, num_proj_shards=1,
               forget_bias=1.0, state_is_tuple=False,
               activation=math_ops.tanh):
    """Initialize the parameters for an LSTM cell.

    Args:
      num_units: int, The number of units in the LSTM cell
      use_peepholes: bool, set True to enable diagonal/peephole connections.
      initializer: (optional) The initializer to use for the weight and
        projection matrices.
      num_proj: (optional) int, The output dimensionality for the projection
        matrices.  If None, no projection is performed.
      proj_clip: (optional) A float value.  If `num_proj > 0` and `proj_clip` is
      provided, then the projected values are clipped elementwise to within
      `[-proj_clip, proj_clip]`.
      num_unit_shards: How to split the weight matrix.  If >1, the weight
        matrix is stored across num_unit_shards.
      num_proj_shards: How to split the projection matrix.  If >1, the
        projection matrix is stored across num_proj_shards.
      forget_bias: Biases of the forget gate are initialized by default to 1
        in order to reduce the scale of forgetting at the beginning of
        the training.
      state_is_tuple: If True, accepted and returned states are 2-tuples of
        the `c_state` and `m_state`.  By default (False), they are concatenated
        along the column axis.  This default behavior will soon be deprecated.
      activation: Activation function of the inner states.
    """
    if not state_is_tuple:
      logging.warn(
          "%s: Using a concatenated state is slower and will soon be "
          "deprecated.  Use state_is_tuple=True." % self)
    self._num_units = num_units
    self._use_peepholes = use_peepholes
    self._initializer = initializer
    self._num_proj = num_proj
    self._proj_clip = proj_clip
    self._num_unit_shards = num_unit_shards
    self._num_proj_shards = num_proj_shards
    self._forget_bias = forget_bias
    self._state_is_tuple = state_is_tuple
    self._activation = activation

    if num_proj:
      self._state_size = (
          rnn_cell.LSTMStateTuple(num_units, num_proj)
          if state_is_tuple else num_units + num_proj)
      self._output_size = num_proj
    else:
      self._state_size = (
          rnn_cell.LSTMStateTuple(num_units, num_units)
          if state_is_tuple else 2 * num_units)
      self._output_size = num_units