RLlib tutorials or courses?

Does anybody know some complete course or tutorial how to use RLlib?
I didn’t find any video tutorial or blog where all details are explained step by step how to set different distributed training options and how practicaly make python code for it.

I have some skills in:

  1. General RL knowledge of algorithms
    1)Keras and tensorflow
    2)basic knowledge of docker and redis
    3)I can write from scratch DDQN or DDPG algorithms

but I don’t know how to combine all these things together.

I decided to use some RL framework because training with single agent instance is too slow and manual control of proceses is to complex.
I see that I don’t understand some general concepts of RLlib because even problem is basic task where my checkpoints are stored and how to use them in RLlib (the goal is not to train but to use trained model)
I would like to train env BipedalWalker


Hi Peter, you can find the official tutorials here.



Thank You, I have started this tutorial and I hope i will understand dependencies between components. Now I improve my skills in redis. I made several agents which are storing own trajectories in redis-ai database I hope it will be training usefull for me.