| Intel(R) Math Kernel Library for Deep Neural Networks (Intel(R) MKL-DNN)
    0.17
    Performance library for Deep Learning | 
An extension for controlling primitive behavior. More...
| Modules | |
| Sequence of post operations | |
| An extension for performing extra operations after base operation. | |
| Functions | |
| mkldnn_status_t MKLDNN_API | mkldnn_primitive_attr_create (mkldnn_primitive_attr_t *attr) | 
| Creates an empty (default) attrattribute.  More... | |
| mkldnn_status_t MKLDNN_API | mkldnn_primitive_attr_clone (mkldnn_primitive_attr_t *attr, const_mkldnn_primitive_attr_t existing_attr) | 
| Makes a copy of an existing_attr.  More... | |
| mkldnn_status_t MKLDNN_API | mkldnn_primitive_attr_destroy (mkldnn_primitive_attr_t attr) | 
| Deletes an attr.  More... | |
| mkldnn_status_t MKLDNN_API | mkldnn_primitive_attr_get_int_output_round_mode (const_mkldnn_primitive_attr_t attr, mkldnn_round_mode_t *round_mode) | 
| Returns integer output rounding mode round_modefor a givenattr, previously set by mkldnn_primitive_attr_set_int_output_round_mode.  More... | |
| mkldnn_status_t MKLDNN_API | mkldnn_primitive_attr_set_int_output_round_mode (mkldnn_primitive_attr_t attr, mkldnn_round_mode_t round_mode) | 
| Sets output rounding mode round_modefor integer operations for a givenattr.  More... | |
| mkldnn_status_t MKLDNN_API | mkldnn_primitive_attr_get_output_scales (const_mkldnn_primitive_attr_t attr, int *count, int *mask, const float **scales) | 
| Returns count, correspondence scalemask, and pointer to a constant floating point array of outputscalesfor givenattr, previously set by mkldnn_primitive_attr_set_output_scales.  More... | |
| mkldnn_status_t MKLDNN_API | mkldnn_primitive_attr_set_output_scales (mkldnn_primitive_attr_t attr, int count, int mask, const float *scales) | 
| Sets output scalesfor primitive operations.  More... | |
| mkldnn_status_t MKLDNN_API | mkldnn_primitive_attr_get_post_ops (const_mkldnn_primitive_attr_t attr, const_mkldnn_post_ops_t *post_ops) | 
| Returns post_opsfor given attr.  More... | |
| mkldnn_status_t MKLDNN_API | mkldnn_primitive_attr_set_post_ops (mkldnn_primitive_attr_t attr, const_mkldnn_post_ops_t post_ops) | 
| Sets configured post_opsto an attributeattrfor future use (when primitive descriptor is being created.  More... | |
An extension for controlling primitive behavior.
| mkldnn_status_t MKLDNN_API mkldnn_primitive_attr_create | ( | mkldnn_primitive_attr_t * | attr | ) | 
Creates an empty (default) attr attribute. 
All the parameters set to default values.
An empty attribute is used in primitive descriptor creating whenever it is not passed explicitly, e.g. in mkldnn_primitive_desc_create.
| mkldnn_status_t MKLDNN_API mkldnn_primitive_attr_clone | ( | mkldnn_primitive_attr_t * | attr, | 
| const_mkldnn_primitive_attr_t | existing_attr | ||
| ) | 
Makes a copy of an existing_attr. 
| mkldnn_status_t MKLDNN_API mkldnn_primitive_attr_destroy | ( | mkldnn_primitive_attr_t | attr | ) | 
Deletes an attr. 
| mkldnn_status_t MKLDNN_API mkldnn_primitive_attr_get_int_output_round_mode | ( | const_mkldnn_primitive_attr_t | attr, | 
| mkldnn_round_mode_t * | round_mode | ||
| ) | 
Returns integer output rounding mode round_mode for a given attr, previously set by mkldnn_primitive_attr_set_int_output_round_mode. 
| mkldnn_status_t MKLDNN_API mkldnn_primitive_attr_set_int_output_round_mode | ( | mkldnn_primitive_attr_t | attr, | 
| mkldnn_round_mode_t | round_mode | ||
| ) | 
Sets output rounding mode round_mode for integer operations for a given attr. 
The default value is mkldnn_round_nearest.
| mkldnn_status_t MKLDNN_API mkldnn_primitive_attr_get_output_scales | ( | const_mkldnn_primitive_attr_t | attr, | 
| int * | count, | ||
| int * | mask, | ||
| const float ** | scales | ||
| ) | 
Returns count, correspondence scale mask, and pointer to a constant floating point array of output scales for given attr, previously set by mkldnn_primitive_attr_set_output_scales. 
scales array points to the internal attr field, so user should not modify/destroy scales.scales is same as attr it belongs to, so it is illegal to use the scales after attr is destroyed | mkldnn_status_t MKLDNN_API mkldnn_primitive_attr_set_output_scales | ( | mkldnn_primitive_attr_t | attr, | 
| int | count, | ||
| int | mask, | ||
| const float * | scales | ||
| ) | 
Sets output scales for primitive operations. 
The number of elements count and correspondence scale mask are stored for future use.
The mask argument defines correspondence between output tensor dimensions and the scales array. Set i-th bit of mask to 1 to use dedicated scaling factor for each slice of the output tensor over i-th dimension. Set mask to 0 to use common scaling factor for the whole output tensor.
Example usage:
count corresponds to mask until an actual primitive descriptor is created, so it is user's responsibility to set proper values. The following formula must be held:
![\[count = \prod\limits_{d \in mask} output.dims[d]\]](form_0.png) 
| mkldnn_status_t MKLDNN_API mkldnn_primitive_attr_get_post_ops | ( | const_mkldnn_primitive_attr_t | attr, | 
| const_mkldnn_post_ops_t * | post_ops | ||
| ) | 
Returns post_ops for given attr. 
post_ops points to the internal attr field, so user should not modify/destroy post_ops. Also the lifetime of post_ops is the same as attr it belongs to, so it is illegal to use post_ops once attr is destroyed. | mkldnn_status_t MKLDNN_API mkldnn_primitive_attr_set_post_ops | ( | mkldnn_primitive_attr_t | attr, | 
| const_mkldnn_post_ops_t | post_ops | ||
| ) | 
Sets configured post_ops to an attribute attr for future use (when primitive descriptor is being created. 
 1.8.13
 1.8.13