NvDsTritonExt#

NVIDIA DeepStream Triton Extension

  • UUID: 68de050b-138f-4f60-af99-e74125ab929e

  • Version: 1.6.0

  • Author: NVIDIA

  • License: Proprietary

Components#

nvidia::deepstream::NvDsTriton#

Nvidia DeepStreamSDK TensorRT Bin. Internal Pipeline: queue->nvinferserver. Refer to nvinferserver plugin

  • Component ID: 93f32aa2-c79c-350c-862f-e090d52f2162

  • Base Type: nvidia::deepstream::INvDsElement

Parameters#

batch-size

Maximum batch size for inference.

  • Flags: GXF_PARAMETER_FLAGS_OPTIONAL

  • Type: GXF_PARAMETER_TYPE_UINT64

  • Default: 1


config-file-path

Path to the configuration file for the current instance of nvinfer.

  • Flags: GXF_PARAMETER_FLAGS_NONE

  • Type: GXF_PARAMETER_TYPE_FILE


filter-out-class-ids

Ignore metadata for objects of specified class ids Use string with values of class ids in ClassID (int) to set the property. e.g. 0;2;3.

  • Flags: GXF_PARAMETER_FLAGS_OPTIONAL

  • Type: GXF_PARAMETER_TYPE_STRING


gpu-id

Set GPU Device ID

  • Flags: GXF_PARAMETER_FLAGS_OPTIONAL

  • Type: GXF_PARAMETER_TYPE_UINT64

  • Default: 0


infer-on-class-ids

Operate on objects with specified class ids Use string with values of class ids in ClassID (int) to set the property. e.g. 0:2:3

  • Flags: GXF_PARAMETER_FLAGS_OPTIONAL

  • Type: GXF_PARAMETER_TYPE_STRING


infer-on-gie-id

Infer on metadata generated by GIE with this unique ID. Set to -1 to infer on all metadata.

  • Flags: GXF_PARAMETER_FLAGS_OPTIONAL

  • Type: GXF_PARAMETER_TYPE_INT64

  • Default: -1


input-tensor-meta

Use preprocessed input tensors attached as metadata instead of preprocessing inside the plugin

  • Flags: GXF_PARAMETER_FLAGS_OPTIONAL

  • Type: GXF_PARAMETER_TYPE_BOOL


interval

Specifies number of consecutive batches to be skipped for inference.

  • Flags: GXF_PARAMETER_FLAGS_OPTIONAL

  • Type: GXF_PARAMETER_TYPE_UINT64

  • Default: 0


model-engine-file

Absolute path to the pre-generated serialized engine file for the model.

  • Flags: GXF_PARAMETER_FLAGS_OPTIONAL

  • Type: GXF_PARAMETER_TYPE_FILE


output-instance-mask

Instance mask expected in network output and attach it to metadata.

  • Flags: GXF_PARAMETER_FLAGS_OPTIONAL

  • Type: GXF_PARAMETER_TYPE_BOOL

  • Default: False


output-tensor-meta

Attach inference tensor outputs as buffer metadata.

  • Flags: GXF_PARAMETER_FLAGS_OPTIONAL

  • Type: GXF_PARAMETER_TYPE_BOOL

  • Default: False


process-mode

Infer processing mode.

Valid values:

1: primary

2: secondary

  • Flags: GXF_PARAMETER_FLAGS_OPTIONAL

  • Type: GXF_PARAMETER_TYPE_INT64

  • Default: 1


raw-output-file-write

Write raw inference output to file.

  • Flags: GXF_PARAMETER_FLAGS_OPTIONAL

  • Type: GXF_PARAMETER_TYPE_BOOL

  • Default: False


unique-id

Unique ID for the element. Can be used to identify output of the element.

  • Flags: GXF_PARAMETER_FLAGS_OPTIONAL

  • Type: GXF_PARAMETER_TYPE_UINT64

  • Default: 15


video-in

Handle to an nvidia::deepstream::NvDsStaticInput component. Supported formats - video(NV12, RGBA).

  • Flags: GXF_PARAMETER_FLAGS_OPTIONAL

  • Type: GXF_PARAMETER_TYPE_HANDLE

  • Handle Type: nvidia::deepstream::NvDsStaticInput


video-out

Handle to an nvidia::deepstream::NvDsStaticOutput component. Supported formats - video(NV12, RGBA).

  • Flags: GXF_PARAMETER_FLAGS_OPTIONAL

  • Type: GXF_PARAMETER_TYPE_HANDLE

  • Handle Type: nvidia::deepstream::NvDsStaticOutput