yolov8安装
- 首先安装conda 使用conda3.8的python虚拟环境创建
- 安装cuda`https://developer.nvidia.com/cuda-toolkit-archive`推荐安装稳定版本
- 安装cudnn
https://developer.nvidia.com/rdp/cudnn-archive - cudnn教程 `https://blog.csdn.net/jhsignal/article/details/111401628`
- 安装pytorch `https://pytorch.org/get-started/locally/`
- clone yolov8项目 `git clone –depth=1 https://gitclone.com/github.com/ultralytics/ultralytics.git`
- 安装 requirements `pip install -r requirements.txt -i https://mirrors.bfsu.edu.cn/pypi/web/simple/`
- 安装
ultralytics 进入此文件夹执行python setup.py install - 下载权重 https://github.com/ultralytics/assets/releases