Intel® oneAPI Deep Neural Network Developer Guide and Reference
enum dnnl::graph::op::attr
Overview
Attributes of operations. More…
#include <dnnl_graph.hpp>
enum attr
{
undef = dnnl_graph_op_attr_undef,
alpha = dnnl_graph_op_attr_alpha,
beta = dnnl_graph_op_attr_beta,
epsilon = dnnl_graph_op_attr_epsilon,
max = dnnl_graph_op_attr_max,
min = dnnl_graph_op_attr_min,
momentum = dnnl_graph_op_attr_momentum,
scales = dnnl_graph_op_attr_scales,
axis = dnnl_graph_op_attr_axis,
begin_norm_axis = dnnl_graph_op_attr_begin_norm_axis,
groups = dnnl_graph_op_attr_groups,
axes = dnnl_graph_op_attr_axes,
dilations = dnnl_graph_op_attr_dilations,
dst_shape = dnnl_graph_op_attr_dst_shape,
kernel = dnnl_graph_op_attr_kernel,
order = dnnl_graph_op_attr_order,
output_padding = dnnl_graph_op_attr_output_padding,
pads_begin = dnnl_graph_op_attr_pads_begin,
pads_end = dnnl_graph_op_attr_pads_end,
shape = dnnl_graph_op_attr_shape,
sizes = dnnl_graph_op_attr_sizes,
src_shape = dnnl_graph_op_attr_src_shape,
strides = dnnl_graph_op_attr_strides,
weights_shape = dnnl_graph_op_attr_weights_shape,
zps = dnnl_graph_op_attr_zps,
group_shape = dnnl_graph_op_attr_group_shape,
exclude_pad = dnnl_graph_op_attr_exclude_pad,
keep_dims = dnnl_graph_op_attr_keep_dims,
keep_stats = dnnl_graph_op_attr_keep_stats,
per_channel_broadcast = dnnl_graph_op_attr_per_channel_broadcast,
special_zero = dnnl_graph_op_attr_special_zero,
transpose_a = dnnl_graph_op_attr_transpose_a,
transpose_b = dnnl_graph_op_attr_transpose_b,
use_affine = dnnl_graph_op_attr_use_affine,
use_dst = dnnl_graph_op_attr_use_dst,
auto_broadcast = dnnl_graph_op_attr_auto_broadcast,
auto_pad = dnnl_graph_op_attr_auto_pad,
coordinate_transformation_mode = dnnl_graph_op_attr_coordinate_transformation_mode,
data_format = dnnl_graph_op_attr_data_format,
mode = dnnl_graph_op_attr_mode,
qtype = dnnl_graph_op_attr_qtype,
rounding_type = dnnl_graph_op_attr_rounding_type,
weights_format = dnnl_graph_op_attr_weights_format,
end = dnnl_graph_op_attr_end,
};Detailed Documentation
Attributes of operations.
Different operations support different attributes. Check the document of each operation for what attributes are supported and what are the potential values for them. Missing required attribute or illegal attribute value may lead to failure when adding the operation to a graph.
Enum Values
undefUndefined op attribute.
alphaSpecifies an alpha attribute to an op.
betaSpecifies an beta attribute to an op.
epsilonSpecifies an epsilon attribute to an op.
maxSpecifies a max attribute to an op.
minSpecifies a min attribute to an op.
momentumSpecifies a momentum attribute to an op.
scalesSpecifies a scales attribute to an op.
axisSpecifies an axis attribute to an op.
begin_norm_axisSpecifies a begin_norm_axis attribute to an op.
groupsSpecifies a groups attribute to an op.
axesSpecifies an axes attribute to an op.
dilationsSpecifies a dilations attribute to an op.
dst_shapeSpecifies an dst_shape attribute to an op.
kernelSpecifies a kernel attribute to an op.
orderSpecifies an order attribute to an op.
output_paddingSpecifies an output_padding attribute to an op.
pads_beginSpecifies a pads_begin attribute to an op.
pads_endSpecifies a pads_end attribute to an op.
shapeSpecifies a shape attribute to an op.
sizesSpecifies a sizes attribute to an op.
src_shapeSpecifies an src_shape attribute to an op.
stridesSpecifies a strides attribute to an op.
weights_shapeSpecifies a weight_shape attribute to an op.
zpsSpecifies a zps attribute to an op.
group_shapeSpecifies the group shape of an op.
The size of the vector should match that of the input. For the dimensions where the grouped quantization occurs, the values should correspond to the group size, which indicates the number of elements that will share the same scaling factor.
exclude_padSpecifies an exclude_pad attribute to an op.
keep_dimsSpecifies a keep_dims attribute to an op.
keep_statsSpecifies a keep_stats attribute to an op.
per_channel_broadcastSpecifies a per_channel_broadcast attribute to an op.
special_zeroSpecifies a special_zero attribute to an op.
transpose_aSpecifies a transpose_a attribute to an op.
transpose_bSpecifies a transpose_b attribute to an op.
use_affineSpecifies an use_affine attribute to an op.
use_dstSpecifies an use_dst attribute to an op.
auto_broadcastSpecifies an auto_broadcast attribute to an op.
The value can be “none” or “numpy”.
auto_padSpecifies an auto_pad attribute to an op.
The value can be “none”, “same_upper”, “same_lower”, or “valid”.
coordinate_transformation_modeSpecifies an coordinate_transformation_mode attribute to an op.
The value can be “half_pixel” or “align_corners”. The attribute is defined for Interpolate operations.
data_formatSpecifies a data_format of an op. The value can be “NCX” or “NXC”.
modeSpecifies a mode attribute of an op.
Interpolate: “nearest”, “linear”, “bilinear”, or “trilinear”. SoftMax: “none”, “inf_as_zero”.
qtypeSpecifies a qtype attribute to an op.
The value can be “per_channel” or “per_tensor”. The attribute is defined for quantization operations.
rounding_typeSpecifies a rounding_type attribute to an op.
The value can be “ceil” or “floor”.
weights_formatSpecifies a weights_format of an op.
The value can be “OIX”, “XIO”, “IOX”, or “XOI”. Different operations may support different values.
endSpecifies the end of all above exteral attributes for check.