RLlib beginners tutorial at this year's Ray Summit (June 22nd-24th)!

Hey everyone, just wanted to share that I’ll be teaching a beginners RLlib tutorial at this year’s Ray Summit (from June 22nd to June 24th). Registration for the summit is free and includes keynotes and many cool talks on the future of distributed computing and AI by industry and research leaders! The tutorials are $50 each. There will be mine on RLlib and another one on Ray Core (taught by my colleague Paco Nathan).

Please use this link to register. Thx! :slight_smile:

6 Likes

Isn’t it free?
I didn’t find any information about the $50 cost.

@Rodney_N_Silva Great question! the Summit is free. There’s an extra cost for the RLlib tutorial on the registration page:

2 Likes

@rliaw can you please share more details about the tutorials, like what will be covered, the duration of the tutorial, as well as whether it is suitable for beginners or somewhat experienced users? Paco Nathan already gave 2 tutorials in the past months, so the community would be better off knowing how this tutorial is going to be different. Thanks

Hey @rfali , great question!
My tutorial will be for beginners and non RL/RLlib experts. It’ll help if you have a background in ML/SL, python, and one of the DL frameworks.
“Hands on RL with RLlib” will consist of 3 x 40min sessions with 15min breaks in between, covering:

  • Intro to RL and RLlib + define our tutorial RL problem (a multi-agent adversarial environment). Exercise at the end.
  • Training with Ray RLlib (initial, “manual” training runs, RLlib default config, where’s what inside a Trainer?, the multi-agent API) and Ray Tune (hyperparameter training). Exercise at the end.
  • Deeper dive: Customizations with RLlib (custom models, custom callbacks). How does RLlib scale? Exercise at the end; time for Q&A.
2 Likes

thanks a lot @sven1977 for the details! Looking forward to it.

1 Like

Will a recording of the tutorial be available to those who are signed up?

Yes, it will be recorded!

Hey everyone, the RLlib tutorial (from last Thursday) is now online for viewing on demand under this link here. You only need to register for the summit to view it:

2 Likes