Antonin RAFFIN
Antonin RAFFIN
Hello, That's true. In practice, this wrapper is not used and the scaling is done during the pre-processing. We may remove that wrapper as it is also now present in...
Related to #60 #70
I did not look closely at the code, but I remember seeing `masks` and `dones` interchangeably in the code (see [here](https://github.com/hill-a/stable-baselines/blob/master/stable_baselines/ppo2/ppo2.py#L383) fir ppo2). Then it used in the [policies](https://github.com/hill-a/stable-baselines/blob/master/stable_baselines/common/policies.py#L262) and...
@HareshMiriyala for now, we don't have time to fix that (even though it is on the roadmap). Currently, we are working on fixing GAIL + A2C, this will be merged...
> It seem the problem is not fixed up to now, is there any plan ? Hello, There is not plan for now, we are now focusing on the migration...
> I think utility functions for reading monitor files would be a better fit yes, we already have that https://stable-baselines3.readthedocs.io/en/master/common/monitor.html#stable_baselines3.common.monitor.load_results >has a messy index column as shown below: What did...
Hello, Do you have a minimal example (minimal custom env) where that error happens? It seems that you are using custom action generation (i.e. not using `.train()`, nor `.predict()` to...
>I guess that's just sth like This works but we need to make an exception for the `_UnVecWrapper`.... cf PR
Hello, Please fill the issue template completely.
>obs = (obs_ph - self.obs_rms.mean) / self.obs_rms.std It looks like this normalization should be deactivated when using discrete observation, otherwise cast error occurs. I'm also not sure if the correct...