Antonin RAFFIN
Antonin RAFFIN
Hello, There is no way currently of doing that and I would definitely appreciate a PR that enables that =). I also tried to fix the seed a while ago...
For the guidelines, please look at the ones in [stable baselines](https://github.com/hill-a/stable-baselines), I use the same here ;)
Hello, are you willing to contribute the implementation?
Probably a new folder would be cleaner.
@corentinlger sorry I was until today at the RL conference, let me try to answer in the coming days when I'm back ;) In short: recurrent PPO in SB3 contrib...
> Here it is the actor and the critic that both incorporate an LSTM component actually, there are different modes in SB3 contrib (shared, actor only, enable critic lstm), the...
> giving the observation and the dones flag to the network. What do you think of this solution ? we need to do that anyway, no?
So, for all people having the issue, please take a look at https://github.com/Stable-Baselines-Team/stable-baselines3-contrib/issues/184#issuecomment-1556859789 in short, if you need to use gym 0.21 (and SB3 v1.x), you need to downgrade both...
Hello, with all those breaking changes, it would maybe make sense to update the package name too? (as we did with SB3) That would avoid many bad surprises when upgrading...
> You might want to check out https://github.com/openai/gym3 the main issue with gym3 (in addition to breaking gym api) is that terminal observation are not handled apparently (see discussion in...