1 安装：
sudo apt-get -y install docker.io
测试：
sudo docker run hello-world
成功：
Hello from Docker!
This message shows that your installation appears to be working correctly.
2 查看
查看已有镜像：
sudo docker images
查看所有容器
sudo docker ps -a
3 下载镜像
下载CUDA docker
1、宿主机需要安装依赖支持CUDA：
distribution=$(. /etc/os-release;echo $ID$VERSION_ID) \&& curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg \&& curl -s -L https://nvidia.github.io/libnvidia-container/$distribution/libnvidia-container.list | \sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' | \sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list
2、安装
sudo apt-get update
sudo apt-get install -y nvidia-docker2
sudo systemctl restart docker
3、dockerhub这个能有nvcc -V，pytorch等安装的cuda都没有nvcc，只有cudatoolkit.
sudo docker pull nvidia/cuda:11.0.3-devel-ubuntu20.04
4 运行
从镜像中创建一个新的容器：
sudo docker run -it --gpus all --name your_container_name your_image_name:v1
文件夹共享：
sudo docker run -it -v /your_dir:/docker_dir --gpus all --name your_container_name your_image_name:v1
启动一个旧的容器:
sudo docker start your_container_name
sudo docker exec -it your_container_name /bin/bash
文件夹拷贝：
docker cp /your_dir your_container_name:/docker_dir
关闭容器、删除容器
sudo docker stop your_container_name
sudo docker rm your_container_name
5 保存容器为镜像，导出加载
保存为文件
sudo docker commit your_container_name your_image_name:v2
导出、加载
docker save -o your_file_name.tar your_image_name:v1
docker load -i your_file_name.tar
