To supoort dockerized JupyterLab on GPU.
$ docker run -d --gpus all \
-p "8888:8888" \
-v notebook:/app/notebook \
--restart always \
--name=cuda-jupyter \
patharanor/cuda-nvidia-jupyter:0.1.0
Before open jupyter-lab via web browser, you need to check jupyter's token first by:
$ docker logs cuda-jupyter
Enjoy on :
YOUR_HOST_IP:8888/?token=TOKEN_ID
The service bind volume from /app/notebook in the container to current directory of host machine ./notebook. So you can get any files via ./notebook.
# Build image
$ docker build -f dev.Dockerfile -t patharanor/cuda-nvidia-jupyter:0.1.0 .
# Push image to DockerHub
$ docker push patharanor/cuda-nvidia-jupyter:0.1.0
MIT
Content type
Image
Digest
Size
4 GB
Last updated
almost 6 years ago
docker pull patharanor/cuda-nvidia-jupyter:0.1.0