NVIDIA-Docker

1、安装nvidia-docker基础镜像（官网拉取，选定自己版本）

docker run --runtime=nvidia --rm nvidia/cuda:9.0-base nvidia-smi
docker run --runtime=nvidia --rm nvidia/cuda:10.0-base nvidia-smi
docker run --runtime=nvidia --rm nvidia/cuda:10.1-base nvidia-smi
docker run --runtime=nvidia --rm nvidia/cuda:11.2-base nvidia-smi

默认下载基础镜像，并启动容器测试，是否成功（安装了nvida-docker2才行，不然会报错没有运行时）（pull手动拉取）

docker pull-nvidia/cuda:9.0-base     

2、查看基础镜像，启动成容器

REPOSITORY        TAG         IMAGE ID        CREATED     SIZE
nvidia/cuda    10.1-base    bfa75f8b799e    3 days ago    105MB

启动容器：sudo docker run -itd --name qinghao nvidia/cuda:10.1-base /bin/bash
进入容器：sudo docker attach qinghao

低版本+需要安装nvidia-docker2才能执行上述命令

docker-19.03及以上版本已经集成了nvidia-docker，但是启动命令变了，需要指定gpu不然不行

docker run  -it --gpus all --rm --name test nvida/cuda:10.1-base /bin/bash

3、换源（/etc/apt/sources.list）

deb http://mirrors.aliyun.com/ubuntu/ xenial main restricted universe multiverse
deb http://mirrors.aliyun.com/ubuntu/ xenial-security main restricted universe multiverse
deb http://mirrors.aliyun.com/ubuntu/ xenial-updates main restricted universe multiverse
deb http://mirrors.aliyun.com/ubuntu/ xenial-proposed main restricted universe multiverse
deb http://mirrors.aliyun.com/ubuntu/ xenial-backports main restricted universe multiverse
deb-src http://mirrors.aliyun.com/ubuntu/ xenial main restricted universe multiverse
deb-src http://mirrors.aliyun.com/ubuntu/ xenial-security main restricted universe multiverse
deb-src http://mirrors.aliyun.com/ubuntu/ xenial-updates main restricted universe multiverse
deb-src http://mirrors.aliyun.com/ubuntu/ xenial-proposed main restricted universe multiverse

deb-src http://mirrors.aliyun.com/ubuntu/ xenial-backports main restricted universe multiverse
deb https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ xenial main restricted universe multiverse
deb https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ xenial-updates main restricted universe multiverse
deb https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ xenial-backports main restricted universe multiverse
deb https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ xenial-security main restricted universe multiverse

deb https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ bionic main restricted universe multiverse
deb-src https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ bionic main restricted universe multiverse
deb https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ bionic-updates main restricted universe multiverse
deb-src https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ bionic-updates main restricted universe multiverse
deb https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ bionic-backports main restricted universe multiverse
deb-src https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ bionic-backports main restricted universe multiverse
deb https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ bionic-security main restricted universe multiverse
deb-src https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ bionic-security main restricted universe multiverse
deb https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ bionic-proposed main restricted universe multiverse
deb-src https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ bionic-proposed main restricted universe multiverse

4、你会发现打不开，vim不好使用。外面新建一个文件，拷进去。

docker cp 外路径 容器:内路径
sudo docker cp sources.list qinghao:/etc/apt/

5、apt update   ---------更新

6、但是会卡壳怎么办？

cd /usr/lib/apt/methods
ln -s http https
apt update

7、安装conda

wget https://repo.continuum.io/archive/Anaconda3-5.2.0-Linux-x86_64.sh
bash Anaconda3-5.2.0-Linux-x86_64.sh

配置系统变量
vim /etc/bash.bashrc

export  PATH="/root/anaconda3/bin:"$PATH

source /etc/bash.bashrc

8、发现没有wget

apt install wget

9、运行程序可以，但是没有中文语言包

apt-get install  language-pack-zh-han* 

export LANGUAGE=zh_CN.UTF-8
export LANG=zh_CN.UTF-8
export LC_ALL=zh_CN.UTF-8

10、缺什么环境就一系列安装就行了

11、容器封装成镜像（下次用这个基镜像FROM，docker build就行了）

sudo docker export qinghao > qinghao-base.tar

12、加载

docker import qinghao-base.tar cuda_python

11&12、是一个错误的步骤

发现不能export和import，但是使用commit可以解决问题

sudo docker commit qinghao  cuda_python

再来一个Dockerfile

FROM cuda_python
ENV PATH /usr/local/bin:$PATH
ENV PATH /root/anaconda3/bin/:$PATH
WORKDIR /code
RUN pip install -i https://pypi.doubanio.com/simple/ --upgrade pip
RUN pip install -i https://pypi.doubanio.com/simple/ tensorflow-gpu==2.2.2
RUN pip install -i https://pypi.doubanio.com/simple/ sklearn
RUN pip install -i https://pypi.doubanio.com/simple/ tqdm
RUN pip install -i https://pypi.doubanio.com/simple/ Keras==2.3.1
RUN pip install -i https://pypi.doubanio.com/simple/ bert4keras==0.10.6

13、可能会出现cudnn不存在，安装一个即可

 conda install -c anaconda cudnn==7.6.4 --yes