tf.contrib.seq2seq.AttentionWrapper

View source on GitHub

Wraps another RNNCell with attention.

Inherits From: RNNCell

tf.contrib.seq2seq.AttentionWrapper(
    cell, attention_mechanism, attention_layer_size=None, alignment_history=False,
    cell_input_fn=None, output_attention=True, initial_cell_state=None, name=None,
    attention_layer=None, attention_fn=None, dtype=None
)

cell An instance of RNNCell.
attention_mechanism A list of AttentionMechanism instances or a single instance.
attention_layer_size A list of Python integers or a single Python integer, the depth of the attention (output) layer(s). If None (default), use the context as attention at each time step. Otherwise, feed the context and cell output into the attention layer to generate attention at each time step. If attention_mechanism is a list, attention_layer_size must be a list of the same length. If attention_layer is set, this must be None. If attention_fn is set, it must guaranteed that the outputs of attention_fn also meet the above requirements.
alignment_history Python boolean, whether to store alignment history from all time steps in the final output state (currently stored as a time major TensorArray on which you must call stack()).
cell_input_fn (optional) A callable. The default is: lambda inputs, attention: array_ops.concat([inputs, attention], -1).
output_attention Python bool. If True (default), the output at each time step is the attention value. This is the behavior of Luong-style attention mechanisms. If False, the output at each time step is the output of cell. This is the behavior of Bhadanau-style attention mechanisms. In both cases, the attention tensor is propagated to the next time step via the state and is used there. This flag only controls whether the attention mechanism is propagated up to the next cell in an RNN stack or to the top RNN output.
initial_cell_state The initial state value to use for the cell when the user calls zero_state(). Note that if this value is provided now, and the user uses a batch_size argument of zero_state which does not match the batch size of initial_cell_state, proper behavior is not guaranteed.
name Name to use when creating ops.
attention_layer A list of tf.compat.v1.layers.Layer instances or a single tf.compat.v1.layers.Layer instance taking the context and cell output as inputs to generate attention at each time step. If None (default), use the context as attention at each time step. If attention_mechanism is a list, attention_layer must be a list of the same length. If attention_layers_size is set, this must be None.
attention_fn An optional callable function that allows users to provide their own customized attention function, which takes input (attention_mechanism, cell_output, attention_state, attention_layer) and outputs (attention, alignments, next_attention_state). If provided, the attention_layer_size should be the size of the outputs of attention_fn.
dtype The cell dtype

TypeError attention_layer_size is not None and (attention_mechanism is a list but attention_layer_size is not; or vice versa).
ValueError if attention_layer_size is not None, attention_mechanism is a list, and its length does not match that of attention_layer_size; if attention_layer_size and attention_layer are set simultaneously.

graph DEPRECATED FUNCTION
output_size Integer or TensorShape: size of outputs produced by this cell.
scope_name

state_size The state_size property of AttentionWrapper.

Methods

get_initial_state

View source

get_initial_state(
    inputs=None, batch_size=None, dtype=None
)

zero_state

View source

zero_state(
    batch_size, dtype
)

Return an initial (zero) state tuple for this AttentionWrapper.

Args
batch_size 0D integer tensor: the batch size.
dtype The internal state data type.

Returns
An AttentionWrapperState tuple containing zeroed out tensors and, possibly, empty TensorArray objects.

Raises
ValueError (or, possibly at runtime, InvalidArgument), if batch_size does not match the output size of the encoder passed to the wrapper object at initialization time.