YOLOv8 YOLOv7 YOLOv5 训练 SCB-Dataset3-U

0 相关资料

YOLOV8环境安装教程.：https://www.bilibili.com/video/BV1dG4y1c7dH/

YOLOV8保姆级教学视频:https://www.bilibili.com/video/BV1qd4y1L7aX/

1 SCB-Dataset3-U 数据

这里使用YOLOv8 训练自定义数据集（SCB-Dataset3）
备注：关于SCB-Dataset3我会公开在github中：
https://github.com/Whiffe/SCB-dataset

在平台中上传数据，通过阿里云盘的方式上传

unzip 5k_HRW_yolo_Dataset.zip
unzip 0.355k_university_yolo_Dataset.zip
unzip 0.71k_university_yolo_Dataset.zip 

2 YOLOv8 训练

2.1 YOLOv8 安装

PyTorch / 2.0.0 / 3.8(ubuntu20.04) / 11.8

git clone https://github.com/ultralytics/ultralytics

推荐

git clone https://gitee.com/YFwinston/ultralytics

cd ultralytics/
pip install ultralytics

2.2 训练的yaml文件

将下面的文件放到：ultralytics/ultralytics/datasets/

0.71k_university_yolo_Dataset.yaml

train: /root/autodl-tmp/0.71k_university_yolo_Dataset/images/train
val: /root/autodl-tmp/0.71k_university_yolo_Dataset/images/val

# number of classes
nc: 6

# class names
names: [ 'hand-raising','reading','writing','using phone','bowing the head','leaning over the table ']

2.3 YOLOv8 训练

下载与上传模型

2.3.1 yolov8n 训练

yolo task=detect mode=train model=yolov8n.pt  data="./ultralytics/datasets/0.71k_university_yolo_Dataset.yaml" batch=8 && /usr/bin/shutdown

2.3.2 yolov8n 验证

yolo task=detect mode=val model=runs/detect/train/weights/best.pt   data="./ultralytics/datasets/0.71k_university_yolo_Dataset.yaml" 

3 YOLOv7 训练

2.2 训练的yaml文件

将下面的文件放到：yolov7/data/

0.71k_university_yolo_Dataset.yaml

train: /root/autodl-tmp/0.71k_university_yolo_Dataset/images/train
val: /root/autodl-tmp/0.71k_university_yolo_Dataset/images/val

# number of classes
nc: 6

# class names
names: [ 'hand-raising','reading','writing','using phone','bowing the head','leaning over the table ']

0.355k_university_yolo_Dataset.yaml

train: /root/autodl-tmp/0.355k_university_yolo_Dataset/images/train
val: /root/autodl-tmp/0.355k_university_yolo_Dataset/images/val

# number of classes
nc: 6

# class names
names: [ 'hand-raising','reading','writing','using phone','bowing the head','leaning over the table ']

YOLOv7x

训练yolov7

python train.py --weights  yolov7x.pt  --data data/0.71k_university_yolo_Dataset.yaml --batch 8 --epochs 100  --cfg ./cfg/training/yolov7x.yaml  && /usr/bin/shutdown

python train.py --weights  yolov7x.pt  --data data/0.355k_university_yolo_Dataset.yaml --batch 8 --epochs 100  --cfg ./cfg/training/yolov7x.yaml  && /usr/bin/shutdown

断点训练

python train.py --weights  runs/train/exp7/weights/last.pt   --data data/0.355k_university_yolo_Dataset.yaml --batch 8 --epochs 100  --cfg ./cfg/training/yolov7x.yaml  --resume && /usr/bin/shutdown

验证yolov7

python test.py --weights runs/train/exp6/weights/best.pt   --data data/0.71k_university_yolo_Dataset.yaml

               Class      Images      Labels           P           R      mAP@.5  mAP@.5:.95: 
                 all         142        4107       0.929        0.75       0.944       0.741
        hand-raising         142           1           1           0       0.995       0.697
             reading         142        1610       0.884       0.851       0.921       0.746
             writing         142         619       0.896       0.859       0.923        0.78
         using phone         142        1470       0.936       0.949       0.969       0.806
     bowing the head         142         185       0.881       0.859       0.873       0.661
leaning over the table          142         222       0.977       0.982       0.985       0.755

python test.py --weights runs/train/exp7/weights/best.pt   --data data/0.355k_university_yolo_Dataset.yaml

               Class      Images      Labels           P           R      mAP@.5  mAP@.5:.95: 
                 all          67        1938       0.911        0.57       0.714       0.551
        hand-raising          67           3           1           0    0.000151    0.000136
             reading          67         782       0.856       0.729       0.819       0.657
             writing          67         329       0.836       0.827       0.883       0.709
         using phone          67         631       0.874       0.868       0.928       0.773
     bowing the head          67          94       0.922       0.501       0.748       0.543
leaning over the table           67          99        0.98       0.497       0.903       0.621