Vanilla Docker Setup for AI/ML in python.
Download and install Docker on your machine: https://docs.docker.com/get-started/.
This should work out of the box if you have a Mac. If you use Windows or Linux, please see the additional instructions below.
It's probably not a good idea to run Docker as root. You can add your username to the docker group by running the following command in the terminal:
sudo usermod -aG docker your-user
You can read more about it here.
If you are running Docker on windows, you will have to enable experimental features. You can find instructions on how to do that here.
Once you have installed Docker installed, use the script provided to run the installation.
$ ./setup_docker.sh
Once the installation is complete, run the Docker machine using the provided script:
$ ./run_docker.sh
If you are running Docker for Windows, you will need to create a code subdirectory where you cloned the repository. For example, say that you cloned the repo to
C:/Users/JohnDoe/AI-ML-Docker-Setup
You will need to create the code subdirectory as follows:
C:/Users/JohnDoe/AI-ML-Docker-Setup/code
You will then need to make the docker daemon aware of this directory to enable the mapping. Instructions for how to do that are found here.
Once you are running the docker image, you are logged into the virtual machine with a bash prompt. Because of a bug in Docker for Mac, you may run into an issue when running jupyter lab from the VM. So, in order to run Jupyter labs you will need to run the following command:
$ jupyter lab --ip=0.0.0.0
Once running, Jupyter Lab will provide you with a URL that looks similar to this:
http://(456937b722b9 or 127.0.0.1):8888/?token=XYZ
Note: Your "token" parameter will be different.
In order to access Jupyter Lab, copy the URL the terminal output, including the token, into your browser on your host machine.
http://127.0.0.1:8888/?token=XYZ
If you would like to view any notebooks that have RISE presentations, unfortunately, you cannot use Jupyter Lab. You must use Jupyter Notebook instead. You can run Jupyter Notebook by issuing the following command:
jupyter notebook --ip=0.0.0.0
and then proceeding as you would with Jupyter Lab.
I have provided a jupyter startup script that will start jupyter notebook without a password. To use it, once you start the VM, execute the following command:
start_jupyter.sh
from the VM command line.
- There are a lot of software packages that will be installed, so be sure to have a fast internet connection and patience..
If the Jupyter notebook is showing a popup showing 403 forbidden errors when trying to create a notebook in the code dir, the permissions need to be changed for the code directory. Try deleting the directory and creating it again using rm -rf code && mkdir code. If that doesn't work, you will need to use sudo chmod 755 code or sudo chown $USER code to set write permissions.