enum dnnl_graph_op_attr_t¶
Overview¶
Attributes of operations. More…
#include <dnnl_graph_types.h> enum dnnl_graph_op_attr_t { dnnl_graph_op_attr_undef = 0, dnnl_graph_op_attr_alpha = 0x1, dnnl_graph_op_attr_beta, dnnl_graph_op_attr_epsilon, dnnl_graph_op_attr_max, dnnl_graph_op_attr_min, dnnl_graph_op_attr_momentum, dnnl_graph_op_attr_scales = 0x20, dnnl_graph_op_attr_axis = 0x30, dnnl_graph_op_attr_begin_norm_axis, dnnl_graph_op_attr_groups, dnnl_graph_op_attr_axes = 0x40, dnnl_graph_op_attr_dilations, dnnl_graph_op_attr_dst_shape, dnnl_graph_op_attr_kernel, dnnl_graph_op_attr_order, dnnl_graph_op_attr_output_padding, dnnl_graph_op_attr_pads_begin, dnnl_graph_op_attr_pads_end, dnnl_graph_op_attr_shape, dnnl_graph_op_attr_sizes, dnnl_graph_op_attr_src_shape, dnnl_graph_op_attr_strides, dnnl_graph_op_attr_weights_shape, dnnl_graph_op_attr_zps, dnnl_graph_op_attr_exclude_pad = 0x60, dnnl_graph_op_attr_keep_dims, dnnl_graph_op_attr_keep_stats, dnnl_graph_op_attr_per_channel_broadcast, dnnl_graph_op_attr_special_zero, dnnl_graph_op_attr_transpose_a, dnnl_graph_op_attr_transpose_b, dnnl_graph_op_attr_use_affine, dnnl_graph_op_attr_use_dst, dnnl_graph_op_attr_auto_broadcast = 0x80, dnnl_graph_op_attr_auto_pad, dnnl_graph_op_attr_coordinate_transformation_mode, dnnl_graph_op_attr_data_format, dnnl_graph_op_attr_mode, dnnl_graph_op_attr_qtype, dnnl_graph_op_attr_rounding_type, dnnl_graph_op_attr_weights_format, dnnl_graph_op_attr_end = 0xFF, };
Detailed Documentation¶
Attributes of operations.
Enum Values¶
dnnl_graph_op_attr_undef
Undefined op attribute.
dnnl_graph_op_attr_alpha
Specifies an alpha attribute to an op.
dnnl_graph_op_attr_beta
Specifies an beta attribute to an op.
dnnl_graph_op_attr_epsilon
Specifies an epsilon attribute to an op.
dnnl_graph_op_attr_max
Specifies a max attribute to an op.
dnnl_graph_op_attr_min
Specifies a min attribute to an op.
dnnl_graph_op_attr_momentum
Specifies a momentum attribute to an op.
dnnl_graph_op_attr_scales
Specifies a scales attribute to an op.
dnnl_graph_op_attr_axis
Specifies an axis attribute to an op.
dnnl_graph_op_attr_begin_norm_axis
Specifies a begin_norm_axis attribute to an op.
dnnl_graph_op_attr_groups
Specifies a groups attribute to an op.
dnnl_graph_op_attr_axes
Specifies an axes attribute to an op.
dnnl_graph_op_attr_dilations
Specifies a dilations attribute to an op.
dnnl_graph_op_attr_dst_shape
Specifies an dst_shape attribute to an op.
dnnl_graph_op_attr_kernel
Specifies a kernel attribute to an op.
dnnl_graph_op_attr_order
Specifies an order attribute to an op.
dnnl_graph_op_attr_output_padding
Specifies an output_padding attribute to an op.
dnnl_graph_op_attr_pads_begin
Specifies a pads_begin attribute to an op.
dnnl_graph_op_attr_pads_end
Specifies a pads_end attribute to an op.
dnnl_graph_op_attr_shape
Specifies a shape attribute to an op.
dnnl_graph_op_attr_sizes
Specifies a sizes attribute to an op.
dnnl_graph_op_attr_src_shape
Specifies a input_shape attribute to an op.
dnnl_graph_op_attr_strides
Specifies a strides attribute to an op.
dnnl_graph_op_attr_weights_shape
Specifies a weight_shape attribute to an op.
dnnl_graph_op_attr_zps
Specifies a zps attribute to an op.
dnnl_graph_op_attr_exclude_pad
Specifies an exclude_pad attribute to an op.
dnnl_graph_op_attr_keep_dims
Specifies a keep_dims attribute to an op.
dnnl_graph_op_attr_keep_stats
Specifies a keep_stats attribute to an op.
dnnl_graph_op_attr_per_channel_broadcast
Specifies a per_channel_broadcast attribute to an op.
dnnl_graph_op_attr_special_zero
Specifies a special_zero attribute to an op.
dnnl_graph_op_attr_transpose_a
Specifies a transpose_a attribute to an op.
dnnl_graph_op_attr_transpose_b
Specifies a transpose_b attribute to an op.
dnnl_graph_op_attr_use_affine
Specifies an use_affine attribute to an op.
dnnl_graph_op_attr_use_dst
Specifies an use_dst attribute to an op.
dnnl_graph_op_attr_auto_broadcast
Specifies an auto_broadcast attribute to an op.
The value can be “none” or “numpy”.
dnnl_graph_op_attr_auto_pad
Specifies an auto_pad attribute to an op.
The value can be “none”, “same_upper”, “same_lower”, or “valid”.
dnnl_graph_op_attr_coordinate_transformation_mode
Specifies an coordinate_transformation_mode attribute to an op.
The value can be “half_pixel” or “align_corners”. The attribute is defined for Interpolate operations.
dnnl_graph_op_attr_data_format
Specifies a data_format of an op. The value can be “NCX” or “NXC”.
dnnl_graph_op_attr_mode
Specifies a mode attribute of an op.
The value can be “nearest”, “linear”, “bilinear”, or “trilinear”. The attribute is defined for Interpolate operations.
dnnl_graph_op_attr_qtype
Specifies a qtype attribute to an op.
The value can be “per_channel” or “per_tensor”. The attribute is defined for quantization operations.
dnnl_graph_op_attr_rounding_type
Specifies a rounding_type attribute to an op.
The value can be “ceil” or “floor”.
dnnl_graph_op_attr_weights_format
Specifies a weights_format of an op.
The value can be “OIX”, “XIO”, “IOX”, or “XOI”. Different operations may support different values.
dnnl_graph_op_attr_end
Specifies the end of all above exteral attributes for check.