Intel® oneAPI Deep Neural Network Developer Guide and Reference
A newer version of this document is available. Customers should click here to go to the newest version.
Visible to Intel only — GUID: GUID-29A3AA19-0A78-408D-A218-508ED3CEAD0A
Visible to Intel only — GUID: GUID-29A3AA19-0A78-408D-A218-508ED3CEAD0A
enum dnnl::algorithm
Overview
Kinds of algorithms. More…
#include <dnnl.hpp>
enum algorithm
{
undef = dnnl_alg_kind_undef,
convolution_auto = dnnl_convolution_auto,
convolution_direct = dnnl_convolution_direct,
convolution_winograd = dnnl_convolution_winograd,
deconvolution_direct = dnnl_deconvolution_direct,
deconvolution_winograd = dnnl_deconvolution_winograd,
eltwise_relu = dnnl_eltwise_relu,
eltwise_tanh = dnnl_eltwise_tanh,
eltwise_elu = dnnl_eltwise_elu,
eltwise_square = dnnl_eltwise_square,
eltwise_abs = dnnl_eltwise_abs,
eltwise_sqrt = dnnl_eltwise_sqrt,
eltwise_swish = dnnl_eltwise_swish,
eltwise_linear = dnnl_eltwise_linear,
eltwise_soft_relu = dnnl_eltwise_soft_relu,
eltwise_mish = dnnl_eltwise_mish,
eltwise_logistic = dnnl_eltwise_logistic,
eltwise_exp = dnnl_eltwise_exp,
eltwise_gelu_tanh = dnnl_eltwise_gelu_tanh,
eltwise_gelu_erf = dnnl_eltwise_gelu_erf,
eltwise_log = dnnl_eltwise_log,
eltwise_clip = dnnl_eltwise_clip,
eltwise_clip_v2 = dnnl_eltwise_clip_v2,
eltwise_pow = dnnl_eltwise_pow,
eltwise_round = dnnl_eltwise_round,
eltwise_hardswish = dnnl_eltwise_hardswish,
eltwise_hardsigmoid = dnnl_eltwise_hardsigmoid,
eltwise_relu_use_dst_for_bwd = dnnl_eltwise_relu_use_dst_for_bwd,
eltwise_tanh_use_dst_for_bwd = dnnl_eltwise_tanh_use_dst_for_bwd,
eltwise_elu_use_dst_for_bwd = dnnl_eltwise_elu_use_dst_for_bwd,
eltwise_sqrt_use_dst_for_bwd = dnnl_eltwise_sqrt_use_dst_for_bwd,
eltwise_logistic_use_dst_for_bwd = dnnl_eltwise_logistic_use_dst_for_bwd,
eltwise_exp_use_dst_for_bwd = dnnl_eltwise_exp_use_dst_for_bwd,
eltwise_clip_v2_use_dst_for_bwd = dnnl_eltwise_clip_v2_use_dst_for_bwd,
lrn_across_channels = dnnl_lrn_across_channels,
lrn_within_channel = dnnl_lrn_within_channel,
pooling_max = dnnl_pooling_max,
pooling_avg_include_padding = dnnl_pooling_avg_include_padding,
pooling_avg_exclude_padding = dnnl_pooling_avg_exclude_padding,
vanilla_rnn = dnnl_vanilla_rnn,
vanilla_lstm = dnnl_vanilla_lstm,
vanilla_gru = dnnl_vanilla_gru,
lbr_gru = dnnl_lbr_gru,
vanilla_augru = dnnl_vanilla_augru,
lbr_augru = dnnl_lbr_augru,
binary_add = dnnl_binary_add,
binary_mul = dnnl_binary_mul,
binary_max = dnnl_binary_max,
binary_min = dnnl_binary_min,
binary_div = dnnl_binary_div,
binary_sub = dnnl_binary_sub,
binary_ge = dnnl_binary_ge,
binary_gt = dnnl_binary_gt,
binary_le = dnnl_binary_le,
binary_lt = dnnl_binary_lt,
binary_eq = dnnl_binary_eq,
binary_ne = dnnl_binary_ne,
resampling_nearest = dnnl_resampling_nearest,
resampling_linear = dnnl_resampling_linear,
reduction_max = dnnl_reduction_max,
reduction_min = dnnl_reduction_min,
reduction_sum = dnnl_reduction_sum,
reduction_mul = dnnl_reduction_mul,
reduction_mean = dnnl_reduction_mean,
reduction_norm_lp_max = dnnl_reduction_norm_lp_max,
reduction_norm_lp_sum = dnnl_reduction_norm_lp_sum,
reduction_norm_lp_power_p_max = dnnl_reduction_norm_lp_power_p_max,
reduction_norm_lp_power_p_sum = dnnl_reduction_norm_lp_power_p_sum,
softmax_accurate = dnnl_softmax_accurate,
softmax_log = dnnl_softmax_log,
};Detailed Documentation
Kinds of algorithms.
Enum Values
undefUndefined algorithm.
convolution_autoConvolution algorithm that is chosen to be either direct or Winograd automatically.
convolution_directDirect convolution.
convolution_winogradWinograd convolution.
deconvolution_directDirect deconvolution.
deconvolution_winogradWinograd deconvolution.
eltwise_reluElementwise: rectified linear unit (ReLU)
eltwise_tanhElementwise: hyperbolic tangent non-linearity (tanh)
eltwise_eluElementwise: exponential linear unit (ELU)
eltwise_squareElementwise: square.
eltwise_absElementwise: abs.
eltwise_sqrtElementwise: square root.
eltwise_swishElementwise: swish (
)
eltwise_linearElementwise: linear.
eltwise_soft_reluElementwise: soft_relu.
eltwise_mishElementwise: mish.
eltwise_logisticElementwise: logistic.
eltwise_expElementwise: exponent.
eltwise_gelu_tanhElementwise: tanh-based gelu.
eltwise_gelu_erfElementwise: erf-based gelu.
eltwise_logElementwise: natural logarithm.
eltwise_clipElementwise: clip.
eltwise_clip_v2Eltwise: clip version 2.
eltwise_powElementwise: pow.
eltwise_roundElementwise: round.
eltwise_hardswishElementwise: hardswish.
eltwise_hardsigmoidElementwise: hardsigmoid.
eltwise_relu_use_dst_for_bwdElementwise: rectified linar unit (ReLU) (dst for backward)
eltwise_tanh_use_dst_for_bwdElementwise: hyperbolic tangent non-linearity (tanh) (dst for backward)
eltwise_elu_use_dst_for_bwdElementwise: exponential linear unit (ELU) (dst for backward)
eltwise_sqrt_use_dst_for_bwdElementwise: square root (dst for backward)
eltwise_logistic_use_dst_for_bwdElementwise: logistic (dst for backward)
eltwise_exp_use_dst_for_bwdElementwise: exponent (dst for backward)
eltwise_clip_v2_use_dst_for_bwdElementwise: clip version 2 (dst for backward)
lrn_across_channelsLocal response normalization (LRN) across multiple channels.
lrn_within_channelLRN within a single channel.
pooling_maxMax pooling.
pooling_avg_include_paddingAverage pooling include padding.
pooling_avg_exclude_paddingAverage pooling exclude padding.
vanilla_rnnRNN cell.
vanilla_lstmLSTM cell.
vanilla_gruGRU cell.
lbr_gruGRU cell with linear before reset.
Differs from the vanilla GRU in how the new memory gate is calculated: 
LRB GRU expects 4 bias tensors on input: 
vanilla_augruAUGRU cell.
lbr_augruAUGRU cell with linear before reset.
binary_addBinary add.
binary_mulBinary mul.
binary_maxBinary max.
binary_minBinary min.
binary_divBinary div.
binary_subBinary sub.
binary_geBinary greater than or equal.
binary_gtBinary greater than.
binary_leBinary less than or equal.
binary_ltBinary less than.
binary_eqBinary equal.
binary_neBinary not equal.
resampling_nearestNearest Neighbor resampling method.
resampling_linearLinear (Bilinear, Trilinear) resampling method.
reduction_maxReduction using max operation.
reduction_minReduction using min operation.
reduction_sumReduction using sum operation.
reduction_mulReduction using mul operation.
reduction_meanReduction using mean operation.
reduction_norm_lp_maxReduction using norm_lp_max operation.
reduction_norm_lp_sumReduction using norm_lp_sum operation.
reduction_norm_lp_power_p_maxReduction using norm_lp_power_p_max operation.
reduction_norm_lp_power_p_sumReduction using norm_lp_power_p_sum operation.
softmax_accurateSoftmax, numerically stable.
softmax_logLogSoftmax, numerically stable.