Ubuntu在线/离线安装nvidia-docker2
版本
ubuntu 18.04
docker 19 (关联)
nvidia-driver 450
cuda 11
1. 在线安装
1.1下载Docker
curl https://get.docker.com | sh \
&& sudo systemctl start docker \
&& sudo systemctl enable docker
1.2 设置NVIDIA Container Toolkit
1.2.1 设置仓库和GPG密钥
distribution=$(. /etc/os-release;echo $ID$VERSION_ID) \
&& curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add - \
&& curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list
1.2.2 更新 apt-get 以及安装nvidia-docker
sudo apt-get update
sudo apt-get install -y nvidia-docker2
sudo systemctl restart docker
1.2.3 测试是否可用 基于CUDA container
sudo docker run --rm --gpus all nvidia/cuda:11.0-base nvidia-smi
结果输出
1.3. 在TensorFlow上使用
下载并运行支持 GPU 的 TensorFlow 映像
docker run --gpus all -it --rm tensorflow/tensorflow:latest-gpu \
python -c "import tensorflow as tf; print(tf.reduce_sum(tf.random.normal([1000, 1000])))"
使用最新的 TensorFlow GPU 映像在容器中启动 bash
shell 会话:
docker run --gpus all -it tensorflow/tensorflow:latest-gpu bash
1.4. Docker 容器换源(可选)
sudo vim /etc/docker/daemon.json
# 插入:
{
"registry-mirrors":["https://1q3uwsii.mirror.aliyuncs.com"]
}
官网链接:
https://www.tensorflow.org/install/docker?hl=zh-cn
https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/install-guide.html#docker
2. 离线安装
2.1 离线安装nvidia-docker2
前提
准备一台可联网的机器 下载相应的包
docker 离线安装 见官网https://docs.docker.com/engine/install/ubuntu/
docker版本>=19
2.1.1 联网机器设置仓库和GPG密钥
distribution=$(. /etc/os-release;echo $ID$VERSION_ID) \
&& curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add - \
&& curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list
$ sudo apt-get update
2.1.2 在联网的机器上下载包
apt download libnvidia-container1
apt download libnvidia-container-tools
apt download nvidia-container-toolkit
apt download nvidia-container-runtime
apt download nvidia-docker2
2.1.3 拷贝到离线机器上安装
# 按顺序执行
dpkg -i libnvidia-container1
dpkg -i libnvidia-container-tools
dpkg -i nvidia-container-toolkit
dpkg -i nvidia-container-runtime
dpkg -i nvidia-docker2
注意:下载的对应版本的包有版本号 敲完命令后tab下补全包的信息
例如: sudo dpkg -i nvidia-docker2_2.5.0-1_all.deb
2.1.4 重启docker服务
systemctl restart docker
感谢:https://blog.csdn.net/Mr_WHITE2/article/details/109532661