Abstract
Object detection is one of the most important areas in computer vision, whichplays a key role in various practical scenarios. Due to limitation of hardware,it is often necessary to sacrifice accuracy to ensure the infer speed of thedetector in practice. Therefore, the balance between effectiveness andefficiency of object detector must be considered. The goal of this paper is toimplement an object detector with relatively balanced effectiveness andefficiency that can be directly applied in actual application scenarios, ratherthan propose a novel detection model. Considering that YOLOv3 has been widelyused in practice, we develop a new object detector based on YOLOv3. We mainlytry to combine various existing tricks that almost not increase the number ofmodel parameters and FLOPs, to achieve the goal of improving the accuracy ofdetector as much as possible while ensuring that the speed is almost unchanged.Since all experiments in this paper are conducted based on PaddlePaddle, wecall it PP-YOLO. By combining multiple tricks, PP-YOLO can achieve a betterbalance between effectiveness (45.% mAP) and efficiency (72.9 FPS), surpassingthe existing state-of-the-art detectors such as EfficientDet and YOLOv4.Sourcecode is at https://github.com/PaddlePaddle/PaddleDetection.