A descriptor for an RNN operation. More...

#include <dnnl_types.h>

Collaboration diagram for dnnl_rnn_desc_t:

[legend]

Public Attributes
dnnl_primitive_kind_t	primitive_kind
	The kind of primitive. More...

dnnl_prop_kind_t	prop_kind
	The kind of propagation. More...

dnnl_alg_kind_t	cell_kind
	RNN cell kind. More...

dnnl_rnn_direction_t	direction
	The direction of RNN primitive execution.

dnnl_memory_desc_t	src_layer_desc
	Source layer memory descriptor.

dnnl_memory_desc_t	src_iter_desc
	Source iteration memory descriptor for hidden state.

dnnl_memory_desc_t	src_iter_c_desc
	Source iteration memory descriptor for cell state.

dnnl_memory_desc_t	weights_layer_desc
	Weights layer memory descriptor.

dnnl_memory_desc_t	weights_iter_desc
	Weights iteration memory descriptor.

dnnl_memory_desc_t	bias_desc
	Bias memory descriptor.

dnnl_memory_desc_t	dst_layer_desc
	Destination layer memory descriptor.

dnnl_memory_desc_t	dst_iter_desc
	Destination iter memory descriptor for hidden state.

dnnl_memory_desc_t	dst_iter_c_desc
	Destination iter memory descriptor for cell state.

dnnl_memory_desc_t	weights_peephole_desc
	Weights peephole memory descriptor This memory descriptor is equal to zero memory descriptor in case of non-peephole LSTMs and other non-LSTM RNNs. More...

dnnl_memory_desc_t	placeholder_desc
	Placeholders.

dnnl_memory_desc_t	diff_src_layer_desc
	Source gradient layer memory descriptor.

dnnl_memory_desc_t	diff_src_iter_desc
	Source gradient iter memory descriptor for hidden state.

dnnl_memory_desc_t	diff_src_iter_c_desc
	Source gradient iter memory descriptor for cell state.

dnnl_memory_desc_t	diff_weights_layer_desc
	Weights gradient layer memory descriptor.

dnnl_memory_desc_t	diff_weights_iter_desc
	Weights gradient iter memory descriptor.

dnnl_memory_desc_t	diff_bias_desc
	Bias gradient memory descriptor.

dnnl_memory_desc_t	diff_dst_layer_desc
	Destination gradient layer memory descriptor.

dnnl_memory_desc_t	diff_dst_iter_desc
	Destination gradient iteration memory descriptor for hidden state.

dnnl_memory_desc_t	diff_dst_iter_c_desc
	Destination gradient iteration memory descriptor for cell state.

dnnl_memory_desc_t	diff_weights_peephole_desc
	Weights gradient peephole memory descriptor This memory descriptor is equal to zero memory descriptor in case of non-peephole LSTMs and other non-LSTM RNNs. More...

dnnl_memory_desc_t	diff_placeholder_desc
	Placeholders.

unsigned int	flags
	RNN cell flags.

dnnl_alg_kind_t	activation_kind
	Activation function used for vanilla_rnn cell kind. More...

Detailed Description

A descriptor for an RNN operation.

Member Data Documentation

◆ primitive_kind

dnnl_primitive_kind_t dnnl_rnn_desc_t::primitive_kind

The kind of primitive.

Used for self-identifying the primitive descriptor. Must be dnnl_rnn.

◆ prop_kind

dnnl_prop_kind_t dnnl_rnn_desc_t::prop_kind

The kind of propagation.

Possible values: dnnl_forward_training, dnnl_forward_inference, and dnnl_backward.

◆ cell_kind

dnnl_alg_kind_t dnnl_rnn_desc_t::cell_kind

RNN cell kind.

Must be one of dnnl_vanilla_rnn, dnnl_vanilla_lstm, dnnl_vanilla_gru, or dnnl_lbr_gru.

◆ weights_peephole_desc

dnnl_memory_desc_t dnnl_rnn_desc_t::weights_peephole_desc

Weights peephole memory descriptor This memory descriptor is equal to zero memory descriptor in case of non-peephole LSTMs and other non-LSTM RNNs.

◆ diff_weights_peephole_desc

dnnl_memory_desc_t dnnl_rnn_desc_t::diff_weights_peephole_desc

Weights gradient peephole memory descriptor This memory descriptor is equal to zero memory descriptor in case of non-peephole LSTMs and other non-LSTM RNNs.

◆ activation_kind

dnnl_alg_kind_t dnnl_rnn_desc_t::activation_kind

Activation function used for vanilla_rnn cell kind.

Must be either dnnl_eltwise_relu or dnnl_eltwise_tanh.

The documentation for this struct was generated from the following file:

include/dnnl_types.h

Public Attributes