# PP-YOLO: An Effective and Efficient Implementation of Object Detector

**Authors:** Xiang Long, Kaipeng Deng, Guanzhong Wang, Yang Zhang, Qingqing Dang, Yuan Gao, Hui Shen, Jianguo Ren, Shumin Han, Errui Ding, Shilei Wen  
**arXiv:** [2007.12099](https://arxiv.org/abs/2007.12099)  
**Date:** 2020-07-23  
**Summary page:** https://www.summarizepaper.com/en/paper/2007.12099v1/

## Key points

- PP-YOLO is an object detector that aims for a balance between effectiveness and efficiency.
- It is based on YOLOv3 and incorporates existing tricks to improve accuracy without significantly increasing model parameters and FLOPs.
- Experiments using PaddlePaddle show better effectiveness (45.% mAP) and efficiency (72.9 FPS) compared to EfficientDet and YOLOv4.
- The source code for PP-YOLO is available on GitHub at https://github.com/PaddlePaddle/PaddleDetection.
- The authors of the paper are Xiang Long, Kaipeng Deng, Guanzhong Wang, Yang Zhang, Qingqing Dang, Yuan Gao, Hui Shen, Jianguo Ren, Shumin Han, Errui Ding, and Shilei Wen.

## Summary

PP-YOLO is an object detector that aims to achieve a balance between effectiveness and efficiency. It is based on the widely used YOLOv3 and combines various existing tricks to improve accuracy while keeping the speed almost unchanged. These tricks do not significantly increase the number of model parameters and FLOPs. All experiments are conducted using PaddlePaddle, resulting in a better balance between effectiveness (45.% mAP) and efficiency (72.9 FPS) compared to state-of-the-art detectors like EfficientDet and YOLOv4. The source code for PP-YOLO is available on GitHub at https://github.com/PaddlePaddle/PaddleDetection. The authors of this paper are Xiang Long, Kaipeng Deng, Guanzhong Wang, Yang Zhang, Qingqing Dang, Yuan Gao, Hui Shen, Jianguo Ren, Shumin Han, Errui Ding and Shilei Wen.