ONE - On-device Neural Engine
|
Class to map parsed tensor and memory for tensor values. For parsed tensor, this memory is used to fill input or output values of graph. More...
#include <TensorDataMap.h>
Public Member Functions | |
TensorDataMap () | |
uint8_t * | allocate (const ParsedTensor *parsed_tensor) |
uint8_t * | data (const ParsedTensor *parsed_tensor) |
Class to map parsed tensor and memory for tensor values. For parsed tensor, this memory is used to fill input or output values of graph.
Definition at line 41 of file TensorDataMap.h.
|
inline |
Definition at line 44 of file TensorDataMap.h.
|
inline |
Definition at line 48 of file TensorDataMap.h.
References nnkit::support::tftestinfo::ParsedTensor::isFloatTensor(), nnkit::support::tftestinfo::ParsedTensor::shape(), and size.
Referenced by nnkit::support::tf::Backend::prepare(), and nnkit::support::tf::Backend::run().
|
inline |
Definition at line 66 of file TensorDataMap.h.
Referenced by nnkit::support::tf::TensorContext::getConstFloatTensor(), nnkit::support::tf::TensorContext::getMutableFloatTensor(), and nnkit::support::tf::Runner::prepareInputs().