YOLOv5绝缘子训练日志

序言

参考链接：https://wanghao.blog.csdn.net/article/details/109253329

https://www.freesion.com/article/5639989910/

https://blog.csdn.net/qq_36756866/article/details/109111065

yolov5链接：https://github.com/ultralytics/YOLOv5

环境

Anaconda=4.9.2

python=3.8.5

keras=2.4.3

tensorflow-gpu=2.3.0

pytorch=1.7.1

numpy=1.20.1

训练

绝缘子图像数量：3410张

单一目标：Insulator

data文件结构：

迭代次数：300次

batch_size=5

weights=weights/yolov5s.pt

耗时：4h52min20s

train训练结果：

 

detect预测结果: