BatchMultiClassNMS

Description

Calculates nms for input boxes and input scores.

Input

Two mandatory inputs and two optional inputs:

boxes: (required) tensor of type float16 or float32.

scores: (required) tensor of type float16 or float32.

clip_window: (optional) tensor of type float16 or float32.

num_valid_boxes: (optional) tensor of type int32.

Output

Four outputs:

nmsed_boxes: tensor of type float16 or float32.

nmsed_scores: tensor of type float16 or float32.

nmsed_classes: tensor of type float16 or float32.

nmsed_num: tensor of type int32.

Attribute

score_threshold: float.

iou_threshold: float.

Max_size_per_class: int type.

max_total_size: int.

change_coordinate_frame: bool.

transpose_box: bool type.

ONNX Opset Support

Opset v8/v9/v10/v11/v12/v13/v14/v15/v16