Hi everyone,
I am training some reinforcement learning models and I am trying to automate the hyperparameter search using optuna. I saw in the documentation that you can use joblib with spark as a backend to train in paralel. I got that working with the example using sklearn, but now that I tried training my model using stable baselines 3 it doesn't seem to be working. Do you know if it is just not possible to do it, or is there a special way to train these models in parallel. I did not want to use Ray yet because SB3 has a lot more models out of the box than RLlib.
Thanks in advance!
