Antonin RAFFIN

hello, > fine-time discretization Could you give a quick example/short explanation of what exact problem it solves that it not solved by other methods?

Thanks a lot @richardjozsa for reviewing =) > calculate the discount for self.n_steps as constant, Good catch, there is an issue with the discount factor used for bootstrapping (`model.gamma` is...

Some benchmark results. Note: n-steps doesn't always improve performance (and usually `n_steps=3` gives equal or better results) ## SAC - LunarLanderContinuous-v3 ``` python train.py --algo sac --env LunarLanderContinuous-v3 -P --n-eval-envs...

Related to https://github.com/DLR-RM/stable-baselines3/pull/1837

Hello, I think you are missing an important alternative, which is also recommended: evaluating each candidate for multiple episodes to remove noise due to env stochasticity. Also, even if you...

Hello, probably related to https://github.com/DLR-RM/stable-baselines3/issues/1500. > This is precisely what sb3 does in their VecFrameStack! In general you should use `VecFrameStack()` instead yes.

> I am wondering why not setting the check to be >=3 instead of strictly equal. https://github.com/DLR-RM/stable-baselines3/blob/fa21bce04ee625c67f6ea2a7678bf46c39cd226c/stable_baselines3/common/preprocessing.py#L35 To me `(2, 3, 64, 64)` looks like a batch of images and...

Hello, this is the expected behavior because `torch.cuda.set_device(3)` was executed in the main thread, whereas the envs are run on other processes. If you want the other processes to use...

> The latest version of rl-baselines3-zoo is 2.3.0, but the tag on docker is still 2.2.0a1 - this updates build_docker.sh so it tags it correctly. Actually this version is for...

fixed in https://github.com/DLR-RM/rl-baselines3-zoo/pull/483/commits/2a7c580d4ec3a5b74e94263210285e1f9d1a555d