Not able to follow yolo_non_max_suppression

My question is related to Autonomous_driving_application_Car_detection and specifically about yolo_non_max_suppression.
What does “filtering_mask = classes == label” mean. And how to go about defining the boxes_label and scores_label and how to use the filtering mask in the definition. I am struggling to start with this code snippet, so any help will be useful.

The purpose of the filtering mask is to select the boxes and scores corresponding to the given label in the current iteration of the loop. They give you some hints in the comments about which TF primitives you need to use to apply the filtering masks to the boxes and scores tensors.