标签：tensor   docke   source   root   ras   get   bss   操作   too   

一，在某目录下有如下文件:

-rw-r--r-- 1 root root 1643293725 9月 2 11:46 cuda_9.0.176_384.81_linux.run
-rw-r--r-- 1 root root 1250 9月 3 17:45 Dockerfile
-rw-r--r-- 1 root root 102521974 9月 2 19:18 libcudnn7_7.0.5.15-1+cuda9.0_amd64.deb
-rwxr-xr-x. 1 root root 111258448 8月 21 16:58 NVIDIA-Linux-x86_64-430.40.run
-rw-r--r-- 1 root root 20766931 8月 30 12:15 Python-3.5.5.tgz
-rw-r--r-- 1 root root 720 9月 3 13:55 sources.list
-rw-r--r-- 1 root root 258835666 9月 3 17:17 tensorflow_gpu-1.11.0-cp35-cp35m-linux_x86_64.whl
二 ，该目录下执行构建命令，时间较长,如：

docker build -t "ubuntu1604-python35-nvidia-cuda90-cudnn7-dockerfil" .

三，使用
nvidia-docker run -it --rm --privileged=true -v $PWD:/data ubuntu1604-python35-nvidia-cuda90-cudnn7-dockerfil /bin/bash

不添加--privileged=true 容器无法正常使用nvidia-smi

Dockerfile 默认无入口命令，即起容器没有后接命令会起不来，如没有 /bin/bash

四，操作目录在 192.168.0.154:/gpu/cuda/1604-py3.5-tensor1.11-keras2.2.4

附：

Dockerfile

FROM ubuntu:16.04
MAINTAINER yon
ENV PATH /usr/local/cuda-9.0/bin:$PATH
ENV LD_LIBRARY_PATH /usr/local/cuda-9.0/lib64:$LD_LIBRARY_PATH
ADD Python-3.5.5.tgz /opt
ADD sources.list /etc/apt/sources.list
RUN apt-get update && apt-get install -y

安装依赖

RUN apt-get install gcc -y && apt-get install make -y && apt-get install vim -y && apt-get install openssl -y && apt-get install libssl-dev -y && apt-get install python3-pip -y
RUN ./opt/Python-3.5.5/configure --prefix=/usr/local/python3.5 && make && make install

COPY libcudnn7_7.0.5.15-1+cuda9.0_amd64.deb /
COPY cuda_9.0.176_384.81_linux.run /
COPY NVIDIA-Linux-x86_64-430.40.run /
COPY tensorflow_gpu-1.11.0-cp35-cp35m-linux_x86_64.whl /
RUN apt-get -y -q install gcc kmod ?
&& sh NVIDIA-Linux-x86_64-430.40.run -a -N --ui=none --no-kernel-module && sh cuda_9.0.176_384.81_linux.run --no-opengl-libs --toolkit -silent && dpkg -i libcudnn7_7.0.5.15-1+cuda9.0_amd64.deb && pip3 install -i https://mirrors.aliyun.com/pypi/simple/ tensorflow_gpu-1.11.0-cp35-cp35m-linux_x86_64.whl && pip3 install -i https://mirrors.aliyun.com/pypi/simple/ Keras==2.2.4

CMD [""]

sources.list

deb http://mirrors.aliyun.com/ubuntu/ xenial main
deb-src http://mirrors.aliyun.com/ubuntu/ xenial main

deb http://mirrors.aliyun.com/ubuntu/ xenial-updates main
deb-src http://mirrors.aliyun.com/ubuntu/ xenial-updates main

deb http://mirrors.aliyun.com/ubuntu/ xenial universe
deb-src http://mirrors.aliyun.com/ubuntu/ xenial universe
deb http://mirrors.aliyun.com/ubuntu/ xenial-updates universe
deb-src http://mirrors.aliyun.com/ubuntu/ xenial-updates universe

deb http://mirrors.aliyun.com/ubuntu/ xenial-security main
deb-src http://mirrors.aliyun.com/ubuntu/ xenial-security main
deb http://mirrors.aliyun.com/ubuntu/ xenial-security universe
deb-src http://mirrors.aliyun.com/ubuntu/ xenial-security universe

ubuntu1604-Python35-cuda9-cudnn7-gpu-dockerfile

标签：tensor   docke   source   root   ras   get   bss   操作   too   

原文地址：https://www.cnblogs.com/g2thend/p/11515569.html