Hi, here’s the link to the blog post: Scaling Deep Reinforcement Learning to a Private Cluster | Stefan’s Blog
Unfortunately, I couldn’t resolve all problems such that it’s still WIP, even though I haven’t worked on this for a while.
Still, I tried to document all problems that I did resolve, so maybe it’s still useful for you.
If you figure out how to solve any of the remaining problems, please let me know - I’m very interested in getting this to work and completing the blog post.