Antonin RAFFIN

Results 880 comments of Antonin RAFFIN

> I don't know how SB3 works internals, would you be able to get one of your devs to update that? Currently, there is only one active dev (me...), Quentin...

Note: the env checker must be updated to warn users that we don't support multi-dim multi discrete and propose a fix (the one from @qgallouedec ).

you may give https://github.com/DLR-RM/stable-baselines3/pull/1837 a try then.

Hello, what is your usecase/expected behavior? the for loop also decompose the info per env: https://github.com/DLR-RM/stable-baselines3/blob/35eccaf04fa011128f02eaecac6caab535686459/stable_baselines3/common/evaluation.py#L99-L106

> A documentation of locals and globals would probably help to find that! :) feel free to open a PR that updates the doc ;)

> Loading the data causes a memory leak through the ep_info_buffer variable. Do you have a minimal example to reproduce/track this behavior? also, how big is the leak?

> Is there anything I should still do for merger? From your side, nothing for now ;) What is missing is from my side. I need to take some time...

I took the time to look at your example closely but I don't understand `model.ep_info_buffer.extend([torch.ones(10000,device="cuda:2")])`, this is not supposed to contain any torch variable. Btw, PPO is usually faster when...

hello, i think there is a misconception between seed, used for pseudo random generator and scenarios.