Non_max_suppression in Tensorflow lite

Hello!
I have YOLOv8 detection model, converted to TensorFlow lite format. Output includes sets of bounding boxes, class id’s and confidence scores and needs some post-processing, include non max suppression. How it can be implemented in TensorFlow Lite for Android device ?

I know that tflite does not support NMS operations so you should try another options to have that results.