miscellane / frames

Infrastructure: Developing Docker Images & Containers for Data Science / Machine Learning

Notes

• Docker
• JupyterLab & Docker
  • Explore
• References

Docker

The directory registry lists a number of docker files, alongside supplements; in continuous development.

JupyterLab & Docker

An interesting article. One of its key examples is

docker container run --rm -i -t -p 10000:8888 jupyter/base-notebook

wherein   $-p \quad 10000:8888$   maps the host port $10000$ to container port $8888$.

Explore

In brief, base image exploration is via

docker run --rm -i -t -p 10000:8888 jupyter/base-notebook

Extend the base image via Dockerfile, and the command

docker build -t notebook .

Subsequently, a container/instance of the image notebook may be used as a development environment via the commands

docker run --rm -i -t -p 127.0.0.1:10000:8888 -w /app --mount type=bind,src="$(pwd)",target=/app notebook

or

docker run -i -t -p 127.0.0.1:10000:8888 -w /app --mount type=bind,src="$(pwd)",target=/app notebook

Note, the container's working environment, i.e., -w, must be inline with this project's top directory. For Visual Studio Code container attachment instructions study Attach Container.

References

• NVIDIA Compute Software Support on Windows Subsystem for Linux (WSL) 2
• Docker & NVIDIA
• conda
• docker hub
• Python Package Index
• pip
  • requirements file format
• jax
  • python package index notes
  • GitHub
  • Documentation