YOLOv8训练自己的目标检测数据集

YOLOv8官方的GitHub代码，同时上面也有基础环境的配置要求以及代码运行的教程。下载后的源码文件名应该是ultralytics–main。

需要额外的安装一些包，因为后面需要用到。

pip install ultralytics

├── linhuo
│ ├── images
│ │ ├── train
│ │ ├── val
│ │ ├── test
│ ├── labels
│ │ ├── train
│ │ ├── v al
│ │ ├── test

数据集文件配置的位置如下：

ultralytics-main/ultralytics/cfg/datasets/

ultralytics-main/ultralytics/cfg/default.yaml

model: './weights/yolov8n.pt'  # (str, optional) path to model file, i.e. yolov8n.pt, yolov8n.yaml

data: 'ultralytics/cfg/datasets/linhuo.yaml'   # (str, optional) path to data file, i.e. coco128.yaml

yolo task=detect mode=train model=yolov8n.pt data=ultralytics/cfg/datasets/linhuo.yaml batch=16 epochs=500

from ultralytics import YOLO
# 权重网络模型加载权重模型
model = YOLO('ultralytics/cfg/models/v8/yolov8.yaml').load('./weights/yolov8n.pt')  # build from YAML and transfer weights

# Train the model: 数据集配置yaml
results = model.train(data='ultralytics/cfg/datasets/linhuo.yaml', epochs=500, batch=16)

from ultralytics import YOLO
model = YOLO('runs/detect/train/weights/best.pt')

# 检测的图片的路径
source = 'ultralytics/assets/2708.jpg'

# 预测结果
results = model.predict(source, save=True)