I wonder that some post about usage ray in docker can be useful. Some people like me, have came to the ray framework not from devops are but from computational science area, so docker useage, cluster configuration and all these stuff is not obvious by the other hand, it’s easier to understand algorithms configurationns, feature engineering etc., so
- Some step by step tutorials how tu use ray (rllib) in docker may be useful. Of course I don’t mean to place docker course here.
- I suggest to add line"COPY Dockerfile /" in Dockerfiles to better understan how its created and modify its.
- Ray-ml tags arn’t precise enouht Docker Hub , many times it needs to pull image and run command “pip list” to show what excatly is inside.
Maybe docker issues need separate forum section ?
I made some simple image to run tensorboard as separate service to visualize ray_results:
https://hub.docker.com/repository/docker/peterpirogtf/tensorboard
Peter