Module: contrib.rnn

Module: tf.contrib.rnn

Module tf.contrib.rnn

Defined in tensorflow/contrib/rnn/__init__.py.

RNN Cells and additional RNN operations.

See RNN and Cells (contrib) guide.

From core

Used to be in core, but kept in contrib.

Created in contrib, eventual plans to move to core.

RNNCell wrappers

RNN functions

Classes

class AttentionCellWrapper: Basic attention cell wrapper.

class BasicLSTMCell: Basic LSTM recurrent network cell.

class BasicRNNCell: The most basic RNN cell.

class BidirectionalGridLSTMCell: Bidirectional GridLstm cell.

class CompiledWrapper: Wraps step execution in an XLA JIT scope.

class CoupledInputForgetGateLSTMCell: Long short-term memory unit (LSTM) recurrent network cell.

class DeviceWrapper: Operator that ensures an RNNCell runs on a particular device.

class DropoutWrapper: Operator adding dropout to inputs and outputs of the given cell.

class EmbeddingWrapper: Operator adding input embedding to the given cell.

class FusedRNNCell: Abstract object representing a fused RNN cell.

class FusedRNNCellAdaptor: This is an adaptor for RNNCell classes to be used with FusedRNNCell.

class GLSTMCell: Group LSTM cell (G-LSTM).

class GRUBlockCell: Block GRU cell implementation.

class GRUCell: Gated Recurrent Unit cell (cf. http://arxiv.org/abs/1406.1078).

class GridLSTMCell: Grid Long short-term memory unit (LSTM) recurrent network cell.

class HighwayWrapper: RNNCell wrapper that adds highway connection on cell input and output.

class InputProjectionWrapper: Operator adding an input projection to the given cell.

class IntersectionRNNCell: Intersection Recurrent Neural Network (+RNN) cell.

class LSTMBlockCell: Basic LSTM recurrent network cell.

class LSTMBlockFusedCell: FusedRNNCell implementation of LSTM.

class LSTMBlockWrapper: This is a helper class that provides housekeeping for LSTM cells.

class LSTMCell: Long short-term memory unit (LSTM) recurrent network cell.

class LSTMStateTuple: Tuple used by LSTM Cells for state_size, zero_state, and output state.

class LayerNormBasicLSTMCell: LSTM unit with layer normalization and recurrent dropout.

class MultiRNNCell: RNN cell composed sequentially of multiple simple cells.

class NASCell: Neural Architecture Search (NAS) recurrent network cell.

class OutputProjectionWrapper: Operator adding an output projection to the given cell.

class PhasedLSTMCell: Phased LSTM recurrent network cell.

class RNNCell: Abstract object representing an RNN cell.

class ResidualWrapper: RNNCell wrapper that ensures cell inputs are added to the outputs.

class TimeFreqLSTMCell: Time-Frequency Long short-term memory unit (LSTM) recurrent network cell.

class TimeReversedFusedRNN: This is an adaptor to time-reverse a FusedRNNCell.

class UGRNNCell: Update Gate Recurrent Neural Network (UGRNN) cell.

Functions

stack_bidirectional_dynamic_rnn(...): Creates a dynamic bidirectional recurrent neural network.

stack_bidirectional_rnn(...): Creates a bidirectional recurrent neural network.

static_bidirectional_rnn(...): Creates a bidirectional recurrent neural network.

static_rnn(...): Creates a recurrent neural network specified by RNNCell cell.

static_state_saving_rnn(...): RNN that accepts a state saver for time-truncated RNN calculation.