RNNSequence#
Versioned name: RNNSequence-5
Category: Sequence processing
Short description: RNNSequence operation represents a series of RNN cells. Each cell is implemented as RNNCell operation.
Detailed description
A single cell in the sequence is implemented in the same way as in RNNCell operation. RNNSequence represents a sequence of RNN cells. The sequence can be connected differently depending on direction attribute that specifies the direction of traversing of input data along sequence dimension or specifies whether it should be a bidirectional sequence. The most of the attributes are in sync with the specification of ONNX RNN operator defined RNNCell.
Attributes
hidden_size
Description: hidden_size specifies hidden state size.
Range of values: a positive integer
Type:
int
Required: yes
activations
Description: activation functions for gates
Range of values: any combination of relu, sigmoid, tanh
Type: a list of strings
Default value: tanh
Required: no
activations_alpha, activations_beta
Description: activations_alpha, activations_beta attributes of functions; applicability and meaning of these attributes depends on chosen activation functions
Range of values: a list of floating-point numbers
Type:
float[]
Default value: None
Required: no
clip
Description: clip specifies bound values [-C, C] for tensor clipping. Clipping is performed before activations.
Range of values: a positive floating-point number
Type:
float
Default value: infinity that means that the clipping is not applied
Required: no
direction
Description: Specify if the RNN is forward, reverse, or bidirectional. If it is one of forward or reverse, then
num_directions = 1
. If it is bidirectional, thennum_directions = 2
. Thisnum_directions
value specifies input/output shape requirements. When the operation is bidirectional, the input goes through forward and reverse ways. The outputs are concatenated.Range of values: forward, reverse, bidirectional
Type:
string
Required: yes
Inputs
1:
X
- 3D tensor of type T1[batch_size, seq_length, input_size]
, input data. It differs from RNNCell 1st input only by additional axis with sizeseq_length
. Required.2:
H
- 3D tensor of type T1[batch_size, num_directions, hidden_size]
, input hidden state data. Required.3:
sequence_lengths
- 1D tensor of type T2[batch_size]
, specifies real sequence lengths for each batch element. In case of negative values in this input, the operation behavior is undefined. Required.4:
W
- 3D tensor of type T1[num_directions, hidden_size, input_size]
, the weights for matrix multiplication. Required.5:
R
- 3D tensor of type T1[num_directions, hidden_size, hidden_size]
, the recurrence weights for matrix multiplication. Required.6:
B
- 2D tensor of type T1[num_directions, hidden_size]
, the sum of biases (weights and recurrence weights). Required.
Outputs
1:
Y
- 4D tensor of type T1[batch_size, num_directions, seq_len, hidden_size]
, concatenation of all the intermediate output values of the hidden.2:
Ho
- 3D tensor of type T1[batch_size, num_directions, hidden_size]
, the last output value of hidden state.
Types
T1: any supported floating-point type.
T2: any supported integer type.
Example
<layer ... type="RNNSequence" ...>
<data hidden_size="128"/>
<input>
<port id="0">
<dim>1</dim>
<dim>4</dim>
<dim>16</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>1</dim>
<dim>128</dim>
</port>
<port id="2">
<dim>1</dim>
</port>
<port id="3">
<dim>1</dim>
<dim>128</dim>
<dim>16</dim>
</port>
<port id="4">
<dim>1</dim>
<dim>128</dim>
<dim>128</dim>
</port>
<port id="5">
<dim>1</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="6">
<dim>1</dim>
<dim>1</dim>
<dim>4</dim>
<dim>128</dim>
</port>
<port id="7">
<dim>1</dim>
<dim>1</dim>
<dim>128</dim>
</port>
</output>
</layer>