Show HN: Setup a gpu optimized spark/hadoop cluster with docker compose

Hey folks, I figured this might be useful for a few people who want to play with hadoop and spark locally. Feedback hugely appreciated.

  I built this from the pain of wanting to setup a cluster for demos. 

I have gpu bits in there due to the use case but figured maybe some could draw inspiration from this.

Happy to answer questions if there's any interest.

Link here: https://github.com/deeplearning4j/docker