kailicht / datascience_toolbox Goto Github PK

Hello, the Dockerfile was one of many I have tried out on GCP to work through the Deeplearning with R book, with the desire for Tensorflow, Keras and RStudio running on a GPU enabled VM.

I tried to start the Dockerfile after building it, and it does not stay up for very long. The logs are unhelpful. I wondered if you would know where I am going wrong? My command to start the container is:

nvidia-docker run --name deeplearning-r --shm-size=1g --ulimit memlock=-1 --ulimit stack=67108864 -e USER=gpu -e PASSWORD=gpu -d -p 80:8787 -v ~/:/home/rstudio gcr.io/gcer-public/rstudio-keras:5b4713e

Any help appreciated :) The Docker image is available publicly if you would like to use it.

So a "~/data/Notebook" directory is created on the host.
Does the symbolic link need to be made inside the container?

It might be worth adding to the docker login that the $ in $oauthtoken has to be escaped, it took me far too long to figure that out, i.e.
docker login -u '$oauthtoken' -p pass123haha nvcr.io