ObjectDetectionOperator Namespace Reference

Classes
class	ObjectDetectionOperator
	Object Detection operator properties for Script Engine . Read the documentation for more info. More...

Enumerations
enum	DetectionState {   NoModel = 0 , Initializing = 1 , InitializingEncrypted = 2 , Loaded = 3 ,   Running = 4 , Stopped = 5 , GeneratingEngineCache = 6 , StoppedUnlicenced = 7 ,   ModelError = 8 , RunError = 9 , InitializationError = 10 , TensorDirNotFound = 11 ,   LicenseError = 12 }
enum	ExecutionProvider { CUDA = 0 , TensorRT = 1 }
enum	TensorRtPrecisionMode { FP32 = 0 , FP16 = 1 }

Enumeration Type Documentation

DetectionState

enum ObjectDetectionOperator.DetectionState

Enumerator
NoModel	0	No model selected.
Initializing	1	Initializing model.
InitializingEncrypted	2	Initializing Composer model.
Loaded	3	Model Loaded.
Running	4	Detection running.
Stopped	5	Operator stopped.
GeneratingEngineCache	6	Generating TensorRT engine cache. This operation might take a few minutes depending on your system and model complexity.
StoppedUnlicenced	7	Operator stopped. 2-hour limit reached for unlicensed Composer model.
ModelError	8	Error loading model (see log file for details)
RunError	9	An Object Detection error occured (see log file for details)
InitializationError	10	Object detection initialization failed (see log file for details)
TensorDirNotFound	11	TensorRT engine cache folder not found. Please verify the path.
LicenseError	12	Licence missing or has expired (see log file for details)

ExecutionProvider

enum ObjectDetectionOperator.ExecutionProvider

Enumerator
CUDA	0	CUDA (default execution provider)
TensorRT	1	TensorRT. Faster inference using optimized engine cache.

TensorRtPrecisionMode

enum ObjectDetectionOperator.TensorRtPrecisionMode

Enumerator
FP32	0	TensorRT Float32 precision mode (default).
FP16	1	TensorRT Float16 precision mode (half precision).