来源：https://blog.csdn.net/sinat_29957455/article/details/124375765
whl和tar.gz文件从这里找：
https://www.cnblogs.com/WenJXUST/p/15481436.html
换源：
(如果正常请不要换源，可以尝试部署Clash到树莓派中，参考：待补充)
sudo nano /etc/apt/sources.list写入下面四个
# 默认注释了源码镜像以提高 apt update 速度，如有需要可自行取消注释
deb https://mirrors.tuna.tsinghua.edu.cn/debian/ bookworm main contrib non-free non-free-firmware
# deb-src https://mirrors.tuna.tsinghua.edu.cn/debian/ bookworm main contrib non-free non-free-firmware
deb https://mirrors.tuna.tsinghua.edu.cn/debian/ bookworm-updates main contrib non-free non-free-firmware
# deb-src https://mirrors.tuna.tsinghua.edu.cn/debian/ bookworm-updates main contrib non-free non-free-firmware
deb https://mirrors.tuna.tsinghua.edu.cn/debian/ bookworm-backports main contrib non-free non-free-firmware
# deb-src https://mirrors.tuna.tsinghua.edu.cn/debian/ bookworm-backports main contrib non-free non-free-firmware
deb https://security.debian.org/debian-security bookworm-security main contrib non-free non-free-firmware # deb-src https://security.debian.org/debian-security bookworm-security main contrib non-free non-free-firmware
另一个：
sudo nano /etc/apt/sources.list.d/raspi.list写入
deb https://mirrors.tuna.tsinghua.edu.cn/raspberrypi/ bookworm main
更新
sudo apt-get update
sudo apt-get upgrade进入conda环境
来自：https://www.bilibili.com/video/BV1Pi421Z7w3
来自：https://milkv.io/zh/docs/duo/application-development/tpu/tpu-yolov5
YOLOv5开发工具包-yolov5-master.zip
https://codeload.github.com/ultralytics/yolov5/zip/refs/heads/master
yolov5n.pt
https://github.com/ultralytics/yolov5/releases/download/v6.2/yolov5n.pt
创建yolo文件夹，将master放进去解压
看到requirements.txt
这个文件里面就是yolov5所需要的环境的版本、所需包、下载命令
cd YOLO/yolov5-master/
pip3 install -r /home/admin/YOLO/yolov5-master/requirements.txt
安装过程很长
如果出现错误：
解决之后再次安装即可，会检查已下载的包名，自动跳过
1.哈希值不匹配
在报错里面会给你好的链接，访问下载
扔到admin里面
cd
pip3 install 文件名
2.奇奇怪怪
找到Collecting后面的就是包名
这里是Collecting MarkupSafe>=2.0 (from jinja2->torch>=1.8.0->
pip3 install MarkupSafe
或者
conda install scipy安装成功
3.去下载whl包离线安装
https://www.cnblogs.com/WenJXUST/p/15481436.html
https://pypi.org/project/tensorflow-cpu/#history
pip install ssssb.whl
4.警告！降低·PYTHON版本：
https://blog.csdn.net/weixin_42888180/article/details/133963348
https://blog.csdn.net/qq_42479695/article/details/121517743
https://blog.csdn.net/weixin_46333910/article/details/126979904
解决方法在另一篇里面
终于...安完了
验证：
python detect.py调用摄像头：
python detect.py --source 0成功
