Makefile that builds Jupyter on top of nvidia/cuda to take advantage of GPU accelleration.
It clones jupyter/docker-stacks, copies Dockerfiles and files referenced by ADD or COPY, and
overrides the BASE_CONTAINER argument in the stack of Dockerfiles.
make all
These images are available from DockerHub:
- pbatey/jupyter-r-notebook-gpu
- pbatey/jupyter-scipy-notebook-gpu
- pbatey/jupyter-datascience-notebook-gpu
- pbatey/jupyter-tensorflow-notebook-gpu
Running these images requires the NVIDIA Container Toolkit 2.0 to be installed. Installalation instructions can be found here.
jupyter/docker-stacks Dockerfiles
Dockerfile (-f) |
Tagged As (-t) |
Base Container (--build-arg BASE_CONTAINER=) |
|---|---|---|
| base-notebook/Dockerfile | pbatey/jupyter-base-notebook-gpu | nvidia/cuda:11.2.2-cudnn8-runtime-ubuntu20.04 |
| minimal-notebook/Dockerfile | pbatey/jupyter-minimal-notebook-gpu | pbatey/jupyter-base-notebook-gpu |
| r-notebook/Dockerfile | pbatey/jupyter-r-notebook-gpu | pbatey/jupyter-minimal-notebook-gpu |
| scipy-notebook/Dockerfile | pbatey/jupyter-scipy-notebook-gpu | pbatey/jupyter-minimal-notebook-gpu |
| datascience-notebook/Dockerfile | pbatey/jupyter-datascience-notebook-gpu | pbatey/jupyter-scipy-notebook-gpu |
| tensorflow-notebook/Dockerfile | pbatey/jupyter-tensorflow-notebook-gpu | pbatey/jupyter-scipy-notebook-gpu |
You can generate useful a Readme for updating docker hub:
scripts/readme.md <type> | pbcopy
Where type is one of r, scipy, datascience, or tensorflow.