doker上运行tensorflow

1>下载

docker pull tensorflow/tensorflow

注:对应的为docker.io/tensorflow/tensorflow

2>运行jupyter

2.1运行

docker run -p 8888:8888 -p 6006:6006 tensorflow/tensorflow

启动后显示信息：

[I 06:36:59.777 NotebookApp] Writing notebook server cookie secret to /root/.local/share/jupyter/runtime/notebook_cookie_secret
[W 06:36:59.816 NotebookApp] WARNING: The notebook server is listening on all IP addresses and not using encryption. This is not recommended.
[I 06:36:59.827 NotebookApp] Serving notebooks from local directory: /notebooks
[I 06:36:59.827 NotebookApp] 0 active kernels
[I 06:36:59.827 NotebookApp] The Jupyter Notebook is running at: http://[all ip addresses on your system]:8888/?token=12d770087a7668b5b0a4aecf12c437069a617bcde42de9b9
[I 06:36:59.827 NotebookApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).
[C 06:36:59.827 NotebookApp]
   
    Copy/paste this URL into your browser when you connect for the first time,
    to login with a token:
        http://localhost:8888/?token=12d770087a7668b5b0a4aecf12c437069a617bcde42de9b9

说明：

1、使用http://localhost:8888/?token=12d770087a7668b5b0a4aecf12c437069a617bcde42de9b9可以进行访问，因为宿主主机进行了映射端口，可以直接在8888上进行操作。

2、启动以后已经是后台服务，关闭终端后，仍然可以进行操作

交互启动

docker run -it -p 8888:8888 -p 6006:6006 tensorflow/tensorflow

后台启动

docker run –d -p 8888:8888 -p 6006:6006 tensorflow/tensorflow

注：启动后在8888端口上监听

2.2通过宿主主机或者本地的浏览器访问tensorflow并测试

版本说明：

见：http://stackoverflow.com/questions/34694701/tensorflow-which-docker-image-to-use

There are four images:

1. b.gcr.io/tensorflow/tensorflow: TensorFlow CPU binary image.
2. b.gcr.io/tensorflow/tensorflow:latest-devel: CPU Binary image plus source code.
3. b.gcr.io/tensorflow/tensorflow:latest-gpu: TensorFlow GPU binary image.
4. gcr.io/tensorflow/tensorflow:latest-devel-gpu: GPU Binary image plus source code.

 

查看容器ip：

[root@bogon ~]# docker inspect 5eb16eb805c3 | grep IPAddress
            "SecondaryIPAddresses": null,
            "IPAddress": "172.17.0.2",
                    "IPAddress": "172.17.0.2",

查看容器端口：

[root@bogon ~]# docker port 5eb16eb805c3
6006/tcp -> 0.0.0.0:6006
8888/tcp -> 0.0.0.0:8888

 

2.3查看相关日志

[root@bogon ~]# docker logs --help

Usage:  docker logs [OPTIONS] CONTAINER

Fetch the logs of a container

  -f, --follow        Follow log output
  --help              Print usage
  --since             Show logs since timestamp
  -t, --timestamps    Show timestamps
  --tail=all          Number of lines to show from the end of the logs

 

docker logs -f f2b4c9bb53dd  

同tail -f的用途

docker logs – t 2b4c9bb53dd

显示日志包括所有的时间戳

 

3>运行TensorBoard

3.1运行

docker run   -p 6006:6006 tensorflow/tensorflow tensorboard --logdir=/opt/tensor

 

3.2浏览器访问：

http://localhost:6006/

 

3.3进入tensorboard容器中运行相关代码

说明：docker中运行外部数据需要挂载外部卷来操作

 

4>启动脚本

#!/bin/bash
setenforce 0
ifconfig | grep -w inet
docker ps -l
echo 按任意键继续
read -n 1
docker run -it  -p 8888:8888 -p 6006:6006 -v /data/article:/article docker.io/tensorflow/tensorflow

绑定一个文件夹启动

 

5>参考文档

scikit.官方例子： http://scikit-learn.org/stable/auto_examples/index.html

TensorBoard简介(docker可以启动可视化界面)  http://www.cnblogs.com/lienhua34/p/5998885.html

tensorflow例子教程 http://wiki.jikexueyuan.com/project/tensorflow-zh/ 

github上tensorflow例子 https://github.com/aymericdamien/TensorFlow-Examples

使用GBDT选取特征    http://www.letiantian.me/2015-03-31-use-gbdt-to-select-features/

scikit官方api  http://scikit-learn.org/stable/

机器学习之用Python从零实现贝叶斯分类器  http://python.jobbole.com/81019/ 

机器学习算法一般步骤 http://www.cnblogs.com/chaoren399/p/4851658.html

基于机器学习方法的POI品类推荐算法   http://tech.meituan.com/category-recommend-base-ml.html

Tensorflow学习笔记3：TensorBoard可视化学习  http://www.cnblogs.com/lienhua34/p/5998885.html