Make use of YOLO-NAS for object detection

We are currently using a very outdated model for object detection (Yolov3). Current state-of-the-art model way outperform that and therefore, we should use them as well. The targeted model here is YOLO-NAS, which can be found here: https://github.com/Deci-AI/super-gradients/blob/master/YOLONAS.md