Ultralytics

YOLOv7

https://github.com/WongKinYiu/yolov7.git

yolov7搭配环境

https://blog.csdn.net/u014613075/article/details/127728506

https://blog.csdn.net/qq_51963216/article/details/125858784

yolov7模型训练结果分析以及如何评估yolov7模型训练的效果

Bookmark: https://blog.csdn.net/weixin_43397302/article/details/127073824

基于YOLOv5的疲劳驾驶检测系统（Python+清新界面+数据集）

https://blog.csdn.net/qq_32892383/article/details/129368345

基于Perclos＆改进YOLOv7的疲劳驾驶DMS检测系统（源码＆教程）

https://blog.csdn.net/m0_73650442/article/details/128001511

YOLOv8

Ultralytics YOLOv8 是一款前沿、最先进（SOTA）的模型，基于先前 YOLO 版本的成功，引入了新功能和改进，进一步提升性能和灵活性。YOLOv8 设计快速、准确且易于使用，使其成为各种物体检测与跟踪、实例分割、图像分类和姿态估计任务的绝佳选择。

先进入自己带pytorch的虚拟环境，与之前的yolo系列安装都不太一样，yolov8仅需要安装ultralytics这一个库就ok了。

pip uninstall ultralytics

pip install ultralytics==8.0.112 -i https://pypi.tuna.tsinghua.edu.cn/simple

CLI指令训练

# 从YAML构建一个新模型，从头开始训练
yolo detect train data=coco128.yaml model=yolov8n.yaml epochs=100 imgsz=640
yolo detect train data=ultralytics/datasets/MoonCake.yaml model=yolov8n.yaml epochs=100 imgsz=640

# 从预训练的*.pt模型开始训练
yolo detect train data=coco128.yaml model=yolov8n.pt epochs=100 imgsz=640
yolo detect train data=ultralytics/datasets/MoonCake.yaml model=yolov8n.pt epochs=100 imgsz=640

# 从YAML中构建一个新模型，将预训练的权重转移到它并开始训练
yolo detect train data=coco128.yaml model=yolov8n.yaml pretrained=yolov8n.pt epochs=100 imgsz=640
yolo detect train data=ultralytics/datasets/MoonCake.yaml model=yolov8n.yaml pretrained=yolov8n.pt epochs=100 imgsz=640

优化器

Bookmark: https://segmentfault.com/a/1190000043540657

参考链接

YOLOV8最强操作教程

https://blog.csdn.net/qq_37706472/article/details/128679699

使用YOLOv8训练自己的【目标检测】数据集

-【收集数据集】-【标注数据集】-【划分数据集】-【配置训练环境】-【训练模型】-【评估模型】-【导出模型】

https://blog.csdn.net/weixin_43694096/article/details/129814232

Yolov8自动标注数据集教程

https://blog.csdn.net/dally2/article/details/131911006

YOLOV8进行疲劳驾驶检测

https://blog.csdn.net/m0_47211450/article/details/130883791

---

优秀网站

• 开源数据集

Bookmark: https://universe.roboflow.com/

Bookmark: https://www.kaggle.com/datasets

• 人脸数据集

Bookmark: http://shuoyang1213.me/WIDERFACE/

• 深度学习可视化工具

Bookmark: https://wandb.ai/site

• yolo

Bookmark: https://docs.ultralytics.com/

• 数据集标注

Bookmark: https://www.makesense.ai/

Bookmark: https://www.robots.ox.ac.uk/~vgg/software/via/

Bookmark: https://yu-xiangdongg.gitee.io/via/via_image_annotator.html

LabelImg标注的YOLO格式txt标签中心坐标和物体边界框长宽的转换

https://blog.csdn.net/Keep_Trying_Go/article/details/128224748

数据集划分

https://blog.csdn.net/weixin_43694096/article/details/129814232

https://blog.csdn.net/weixin_45921929/article/details/128673338

• 神经网络结构可视化软件Netron

Bookmark: https://netron.app/

OpenSeeFace

Bookmark: https://github.com/emilianavt/OpenSeeFace/tree/master

onnxruntime

算力云

Bookmark: https://featurize.cn/

Bookmark: https://www.autodl.com/home

大模型

https://llama.meta.com/llama2