EfficientNMS_TRT#

EfficientNMS_TRT - 1#

Version

  • name: EfficientNMS_TRT (GitHub)

  • domain: main

  • since_version: 1

  • function:

  • support_level: SupportType.COMMON

  • shape inference: True

This version of the operator has been available since version 1.

Summary

Attributes

  • background_class - INT (required) : Background class ID.

  • box_coding - INT (required) : Encoding type for the boxes or anchors inputs.

  • iou_threshold - FLOAT (required) : Box IOU threshold value.

  • max_output_boxes - INT (required) : Max detections to output.

  • plugin_version - STRING (required) : Version number of the TRT plugin.

  • score_activation - INT (required) : Activation function to apply to the scores input.

  • score_threshold - FLOAT (required) : Score threshold value.

Inputs

Between 2 and 3 inputs.

  • boxes (heterogeneous) - T:

  • scores (heterogeneous) - T:

  • anchors (optional, heterogeneous) - T:

Outputs

  • num_detections (heterogeneous) - tensor(int32):

  • detection_boxes (heterogeneous) - T:

  • detection_scores (heterogeneous) - T:

  • detection_classes (heterogeneous) - tensor(int32):

Type Constraints

  • T in ( tensor(float), tensor(float16) ): Constrain input and output types to float tensors.

Examples