To run Jupyter Noteboook with CUDA:
- Execute on the command line:
docker run --gpus all --rm -i -t -p 8888:8888 openiss-cuda-conda-jupyter:F22-01 /bin/bash -c "/opt/conda/bin/jupyter notebook --notebook-dir=/opt/notebooks --ip='*' --port=8888 --no-browser --allow-root"
- Open browser with the IPv4 link printed on the conosole to login to Juypter notebook
- Load your notebook
- While running if need to install packages in the container open terminal from Docker Desktop and use the standard commands to install packages.
Resources:
- Docker file:
Dockerfile.jupyter-conda - Binary image on the dockerhub: https://hub.docker.com/r/openiss/openiss-cuda-conda-jupyter/tags
To run:
- Execute:
docker run --name oi-reid --gpus all --rm -e DISPLAY=<your_ipv4_address>:0.0 openiss-reid:F22-01
- Init and activate conda enviroment inside the container as needed
Resources:
- Docker file:
Dockerfile.oi-reid - Binary image on the dockerhub: https://hub.docker.com/r/openiss/openiss-reid/tags