TensorRT Inference API

Detailed Description

The GIE_Context class provides interfaces to load the Caffe model and perform inference with the TensorRT model.

Data Structures
class	Logger
struct	Profiler
class	GIE_Context

Macros
#define	CHECK(status)

Variables
static const int	TIMING_ITERATIONS = 1000
static const int	NUM_BINDINGS = 3

Macro Definition Documentation

#define CHECK

(

status

)

Value:

{                                                       \
    if (status != 0)                                    \
    {                                                   \
        std::cout << "Cuda failure: " << status;        \
        abort();                                        \
    }                                                   \
}

Definition at line 73 of file gie_inference.h.

Variable Documentation

const int NUM_BINDINGS = 3

static

Definition at line 83 of file gie_inference.h.

const int TIMING_ITERATIONS = 1000

static

Definition at line 82 of file gie_inference.h.

Referenced by Profiler::printLayerTimes().