几篇文章保存下：

１．https://blog.csdn.net/guleileo/article/details/80581858

（转于https://blog.csdn.net/App_12062011/article/details/77554288）

２．https://blog.csdn.net/qq_28123095/article/details/80052308

https://zhuanlan.zhihu.com/p/25236464

３．https://www.cnblogs.com/fariver/p/7446921.html

４．https://blog.csdn.net/jesse_mx/article/details/53925356

５.超级好的使用教程;https://github.com/AlexeyAB/darknet#how-to-train-to-detect-your-custom-objects

6.https://www.jianshu.com/p/c1a009385f59

7.https://blog.csdn.net/u014380165/article/details/80202337

8.https://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650741259&idx=1&sn=03dfb0fa3396e5464fc358b5a803e7bf&chksm=871ade75b06d5763a45f3c5da1ca62023a13c5cf7ce52a0a23e6c320f129f79bb9b3be4d2da0&scene=21#wechat_redirect

论文中提到YOLOv3在AP50和小目标上表现都不错，但AP75表现就比较乏力。说明随着IoU的升高，YOLOv3预测的bbox不能很好的与GT相重叠。 

9.https://github.com/ayooshkathuria/pytorch-yolo-v3

以及从该github上扩展出来的

检测：https://blog.paperspace.com/how-to-implement-a-yolo-object-detector-in-pytorch/

训练：https://github.com/marvis/pytorch-yolo2