File QnnHtaBackend.h¶
↰ Parent directory (include/QNN/HTA)
QNN HTA component Backend API.
Contents
Definition (include/QNN/HTA/QnnHtaBackend.h)¶
Detailed Description¶
The interfaces in this file work with the top level QNN API and supplements QnnBackend.h for HTA backend
Includes¶
QnnBackend.h(File QnnBackend.h)
Full File Listing¶
QNN HTA component Backend API.
The interfaces in this file work with the top level QNN
API and supplements QnnBackend.h for HTA backend
Defines
-
QNN_HTA_BACKEND_CUSTOM_CONFIG_INIT { 0 /*bitmaskFeatures*/ }
QnnHtaBackend_CustomConfig_t initializer macro.
Enums
-
enum QnnHtaBackend_Features_t¶
Values:
-
enumerator QNN_HTA_FOLD_RELU_ACTIVATION_INTO_CONV_OFF = 1 << 0¶
The accelerator will always attempt to fold relu activation into the immediate preceding convolution operation. This optimization is correct when quantization ranges for convolution are equal or subset of the Relu operation. For graphs, where this cannot be guranteed, the client should set this flag
-
enumerator QNN_HTA_BACKEND_FEATURES_UNKNOWN = 0x7fffffff¶
UNKNOWN enum event that must not be used.
-
enumerator QNN_HTA_FOLD_RELU_ACTIVATION_INTO_CONV_OFF = 1 << 0¶
-
struct QnnHtaBackend_CustomConfig_t¶
- #include <QnnHtaBackend.h>
Structure describing the set of configurations supported by the backend. Objects of this type are to be referenced through QnnBackend_CustomConfig_t.
Public Members
-
uint32_t bitmaskFeatures¶
field to save the features that are passed via QnnHtaBackend_Features_t
-
uint32_t bitmaskFeatures¶