# 官网地址

 https://docs.ultralytics.com/usage/cli/
 https://kkgithub.com/ultralytics/ultralytics/blob/main/README.zh-CN.md   # github

安装 YOLO

 #pip 安装YOLO （推荐）
 pip install ultralytics    

CLI 使用

1、语法

 yolo TASK MODE ARGS
 #通过   yolo  help 查看更加详细的参数
 TASK (optional) is one of {'segment', 'detect', 'classify', 'obb', 'pose'}
 MODE (required) is one of {'benchmark', 'predict', 'track', 'export', 'val', 'train'}
 ARGS (optional) are any number of custom 'arg=value' pairs like 'imgsz=320' that override defaults.
 #示例：
 yolo train data=coco8.yaml model=yolo11n.pt epochs=10 lr0=0.01

2、Supported Tasks and Modes

YOLO11 builds upon the versatile model range introduced in YOLOv8, offering enhanced support across various computer vision tasks:

3、入门示例

 cd  /home/bbx/software/yolo
 yolo predict model=yolo11n.pt source='https://kkgithub.com/ultralytics/assets/releases/download/v0.0.0/bus.jpg'
 #注：如果模型无法下载， 请手工下载后放在 /home/bbx/software/yolo/modles 目录内或者将 “/home/bbx/.local/lib/python3.8/site-packages/ultralytics/utils/downloads.py” 中github 地址修改为kkgithub
 #如果是手工下载的模型，请执行
 yolo predict model=./modles/yolo11n.pt source='https://kkgithub.com/ultralytics/assets/releases/download/v0.0.0/bus.jpg'

4、模型训练

 cd  /home/bbx/software/yolo
 yolo detect train data=coco.yaml model=./modles/yolo11n.pt epochs=100 imgsz=640

 # coco.yaml 文件内容如下
 # Ultralytics YOLO <d83d><de80>, AGPL-3.0 license
 # COCO8 dataset (first 8 images from COCO train2017) by Ultralytics
 # Documentation: https://docs.ultralytics.com/datasets/detect/coco8/
 # Example usage: yolo train data=coco8.yaml
 # parent
 # ├── ultralytics
 # └── datasets
 #     └── coco8  ← downloads here (1 MB)
 
 # Train/val/test sets as 1) dir: path/to/imgs, 2) file: path/to/imgs.txt, or 3) list: [path/to/imgs1, path/to/imgs2, ..]
 path: ./detect # dataset root dir
 train: images/train # train images (relative to 'path') 
 val: images/val # val images (relative to 'path')
 test: # test images (optional)
 
 # Classes
 names:
   0: seal

5、模型格式转换--需要时转换

 cd /home/bbx/software/yolo
 yolo export model= ./runs/detect/train/weights/best.pt format=onnx   # 导出的模型存放于     /home/bbx/software/yolo/runs/detect/train/weights 目录下

6、模型推理

• 推理相关的官方文档

https://docs.ultralytics.com/modes/predict/

• 命令行推理

yolo detect predict   model=/home/bbx/software/yolo/dataset/runs/detect/train2/weights/best.pt   source='/home/bbx/software/jetson-inference/python/training/detection/ssd/data/hg2/test/*.jpg'

• python 脚本推理

from ultralytics import YOLO

# Load Yolo model
model = YOLO("/home/bbx/software/yolo/dataset/runs/detect/train2/weights/best.pt")
# run inference 
results =  model("/home/bbx/software/jetson-inference/python/training/detection/ssd/data/hg2/test/*.jpg",save=True ,show=False ,show_conf=True, stream=True)
#view  results
for r in results:
    print("~~~~~~~~~~~~~~~~~~~~~~~~~~~~")
    print(r.boxes.conf)   # 置信度
    print(r.boxes.cls)    # 分类ID
    print(r.to_json())
    print(r.names)        #分类名称

Tags：anyscript

微信扫一扫：分享

微信里点“发现”，扫一下

二维码便可将本文分享至朋友圈。