tsmd44/notebook repository overview

⁠notebook

This image is based on tensorflow-notebook⁠ based on ubuntu:latest⁠.

Additional OS packages

language-pack-ja
mecab
libmecab-dev
mecab-ipadic-utf8
curl
file
p7zip-full
graphviz
mysql-client
libmysqlclient-dev
crf++
cabocha

Additional python packages

mysqlclient
psycopg2
graphviz
pydot
faker
opencv-python
mecab-python3
emoji
gensim
nltk
music21
pydub
babel
pymc3
pytorch
torchvision
fastText
cabocha

⁠Jupyter options

You can run the container with jupyter notebook options.

docker run -p 8888:8888 tsmd44/notebook start-notebook.sh --NotebookApp.notebook_dir='/home/jovyan'

You can also directly change the configuration file inside a container.

/home/jovyan/.jupyter/jupyter_notebook_config.py

A full set of options is available here⁠.

⁠Docker optinos

• -e TZ=Asia/Tokyo - Change timezone.
• -e GRANT_SUDO=yes --user root - This option is useful when you wish to install OS packages with apt or modify root-owned files. (You don't have to be root when use pip or conda)
• -v /some/host/directory:/home/jovyan/work - Mount some host directory on the container working directly.

Further information is available here⁠.

⁠Password setup

Inside a container, run the following command.

python -c 'from notebook.auth import passwd; passwd()'

Above command will promote you to enter the password twice, and then print the hashed password.

Create container with that hashed password.

docker run -d -p 8888:8888 tsmd44/notebook start-notebook.sh --NotebookApp.password='sha1:xxxxxxxxxxxx:xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx'

Or set the hashed password to /home/jovyan/.jupyter/jupyter_notebook_config.py.

c.NotebookApp.password = 'sha1:xxxxxxxxxxxx:xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx'