0
$\begingroup$

I read most of offline algorithms from Stable Baselines 3 should not be subprocessed (see here : https://stable-baselines3.readthedocs.io/en/master/guide/rl_tips.html - "Which algorithm should I use?" section).

I currently use SAC with a SubprocVecEnv wrapper and it seems work except episode based "train_freq" (ex : (10, 'episode')) cannot be used...

Is usage of multiprocess have a kind of "conceptual incompatibility" with offline algorithm or is it a problem of SB' documentation updating ?

$\endgroup$

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.