b-it-bots Docker images

You can easily create all images using docker-compose:

docker-compose build

GPU Notebook

This image contains several packages for deep learning projects with NVidia GPU support.

• Build this image

  The following example uses base image from a2s-institue gpu-notebook image

  bash build_and_publish.sh --base-image ghcr.io/a2s-institute/docker-stacks/gpu-notebook \
  --base-image-tag 11.8.0-cudnn8-runtime-ubuntu22.04 --registry ghcr.io --publish "" \
  --image gpu-notebook
  

  You can build this image using different cuda versions available here.

• Run the image locally

  docker run --gpus all --name gpu-notebook -it --rm -d -p 8880:8888 ghcr.io/b-it-bots/docker/gpu-notebook:11.8.0-cudnn8-runtime-ubuntu22.04
  

• Login to the container

  docker exec -ti gpu-notebook bash
  
  # check nvidia
  nvidia-smi