I’ve started Stanford class CS224n: Natural Language Processing with Deep Learning
Part of this is to setup and runnig a Jupyter Notebook; but I dont want to install anything locally, in the end, docker should be the tool to run all this things.
Part of the learning points I have to myself is to adquire how to scale any necesary element if I need it.
Run Jupyter Notebook in Docker
This is the
docker-compose.yml to run one notebook:
version: '3' services: datascience-notebook: image: jupyter/datascience-notebook volumes: - ./:/home/jovyan/work ports: - 8888:8888 container_name: jupyter_notebook command: "start-notebook.sh"
./ sets a folder where you will access your files inside jupyter
When you start with
docker-compose up will show a message like:
jupyter_notebook | Executing the command: jupyter notebook jupyter_notebook | [I 23:37:08.146 NotebookApp] Writing notebook server cookie secret to /home/jovyan/.local/share/jupyter/runtime/notebook_cookie_secret jupyter_notebook | [I 23:37:09.002 NotebookApp] JupyterLab extension loaded from /opt/conda/lib/python3.7/site-packages/jupyterlab jupyter_notebook | [I 23:37:09.003 NotebookApp] JupyterLab application directory is /opt/conda/share/jupyter/lab jupyter_notebook | [I 23:37:09.006 NotebookApp] Serving notebooks from local directory: /home/jovyan jupyter_notebook | [I 23:37:09.006 NotebookApp] The Jupyter Notebook is running at: jupyter_notebook | [I 23:37:09.007 NotebookApp] http://7ab23d98d2b8:8888/?token=8eb71442202c5bcfb43c06e7a82e619da72f3c1161a02128 jupyter_notebook | [I 23:37:09.007 NotebookApp] or http://127.0.0.1:8888/?token=8eb71442202c5bcfb43c06e7a82e619da72f3c1161a02128 jupyter_notebook | [I 23:37:09.007 NotebookApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation). jupyter_notebook | [C 23:37:09.011 NotebookApp] jupyter_notebook | jupyter_notebook | To access the notebook, open this file in a browser: jupyter_notebook | file:///home/jovyan/.local/share/jupyter/runtime/nbserver-6-open.html jupyter_notebook | Or copy and paste one of these URLs: jupyter_notebook | http://7ab23d98d2b8:8888/?token=8eb71442202c5bcfb43c06e7a82e619da72f3c1161a02128 jupyter_notebook | or http://127.0.0.1:8888/?token=8eb71442202c5bcfb43c06e7a82e619da72f3c1161a02128
To open the jupyter notebooks on you local browser you need to check your
docker-machine IP and replace it into the corresnpodent url: