GPU-enabled docker container with Jupyterlab for artificial intelligence

Published article: https://doi.org/10.1093/gigascience/giad028

General information

Project name: An accessible infrastructure for artificial intelligence using a docker-based Jupyterlab in Galaxy

Project home page: https://github.com/usegalaxy-eu/gpu-jupyterlab-docker

Docker file: https://github.com/usegalaxy-eu/gpu-jupyterlab-docker/blob/master/Dockerfile

Container at Quay.io: https://quay.io/galaxy/docker-ml-jupyterlab/tags

Galaxy tool (that runs this container): https://github.com/usegalaxy-eu/galaxy/blob/release_23.0_europe/tools/interactive/interactivetool_ml_jupyter_notebook.xml

Data: https://zenodo.org/record/6091361 (to run sample notebooks at https://github.com/anuprulez/gpu_jupyterlab_ct_image_segmentation)

How to use: Galaxy training network tutorial

Operating system(s): Linux

Programming language(s): Python, Docker, XML

iPython sample notebooks: https://github.com/anuprulez/gpu_jupyterlab_ct_image_segmentation

Other requirements: Docker 20.10.21, (Optional) CUDA 11.8, CUDA DNN 8

License: MIT License

RRID: SCR_022695

bioToolsID: gpu-enabled_docker_container_with_jupyterlab_for_ai

Running steps:

1. Download container: docker pull quay.io/galaxy/docker-ml-jupyterlab:galaxy-integration-0.3

2. Run container (on host without Nvidia GPU): docker run -it -p 8888:8888 -v <<path to local folder>>:/import quay.io/galaxy/docker-ml-jupyterlab:galaxy-integration-0.3

3. Run container (on host with Nvidia GPU): docker run -it --gpus all -p 8888:8888 -v <<path to local folder>>:/import quay.io/galaxy/docker-ml-jupyterlab:galaxy-integration-0.3

4. Open the link to the Jupyterlab (e.g. http://<<host>>:8888/ipython/lab)

List of packages for Machine learning and deep learning

• Python (version: 3.10.0)
• Jupyterlab (version: 3.6.5)
• Jupyterlab-git (version: 0.41.0)
• Scikit learn (version: 1.1.2)
• Scikit image (version: 0.21.0)
• Tensorflow (version: 2.11)
• ONNX (version: 1.12.0)
• Nibabel (5.1.0)
• OpenCV (version: 4.7)
• CUDA (version: 11.8)
• CUDA DNN (version: 8.6)
• Bqplot (version: 0.12.39)
• Bokeh (version: 3.2.0)
• Matplotlib (version: 3.7.2)
• Seaborn (version: 0.12.2)
• Voila (version: 0.4.1)
• Jupyterlab-nvdashboard (version: 0.8.0)
• Py3Dmol (version: 2.0.3)
• Elyra AI (version: 3.15.0)
• Colabfold (version: 1.5.2)
• Bioblend (version: 1.1.1)
• Biopython (version: 1.81)
• many more ...