NotePublic/Software/Applications/Tensorflow/Docker_TensorFlow_Jupyter_环境搭建.md

Docker+TensorFlow+Jupyter 环境搭建

1.安装和配置

# Tensorflow1
docker run -t -p 1008:8888 -e JUPYTER_ENABLE_LAB=yes --name=tensorflow --privileged -v <host dir>:/home/jovyan jupyter/tensorflow-notebook:17aba6048f44
# Tensorflow2(ubuntu-18.04)
docker run -t -p 1008:8888 -e JUPYTER_ENABLE_LAB=yes --name=tensorflow --privileged -v <host dir>:/home/jovyan jupyter/tensorflow-notebook:ubuntu-18.04

之后通过浏览器访问对应 IP:Port 即可。

2.使能 Jupyter Lab 插件

settings->Advanced Settings Editor->Extension Manager

在 User Overrides 中进行修改

{
    "enabled": true
}

然后 Save User Settings，刷新网页。侧边栏会显示“插件选项”在其中可进行搜索、安装和卸载等插件管理操作。

3.Jupyter Notebook 配置

docker exec -it <container name/id> pip install jupyter_contrib_nbextensions
docker exec -it <container name/id> jupyter contrib nbextension install --user --skip-running-check
# 重启容器
docker stop <container name/id>
docker start -a <container name/id>

访问 Jupyter Notebook->Nbextensions，勾选：

1. Highlight selected word
2. highlighter
3. jupyter-js-widgets/extension
4. Table of Contents
5. contrib_nbextensions_help_item
6. Hinterland
7. jupyter-matplotlib/extension
8. Nbextensions dashboard tab
9. Nbextensions edit menu item