Markus Krimmel

Questions regarding mujoban environment

3

I have two questions regarding the mujoban environment: ### Warnings and errors Let us set up the environment as explained in the README and let's perform some random actions: ```...

Changing `Model.sense` makes (feasible) relaxation infeasible

13

I am trying to determine whether the feasibility region of an MIP-relaxation is bounded or not. To do that, I pick a random objective and both minimize and maximize it,...

It seems like `GymWrapper` does not fully comply with the 0.22.0 Gym API (which is what is stated in requirements.txt). There were lots of subtle changes in the Gym API...

- [x] I have marked all applicable categories: + [x] exception-raising bug + [ ] RL algorithm bug + [ ] documentation request (i.e. "X is missing from the documentation.")...

Seeding in 'reset_game'

1

I have a question regarding the method `reset_game` in Base_Agent. The first few lines read: ``` def reset_game(self):Calling seed """Resets the game information so we are ready to play a...

Conflicting Dependencies

2

Installing the repository on Linux (as detailed in the README) gives me the following error when executing `pip install -e .`: ``` ERROR: Cannot install h-baselines because these package versions...

fix: make `pytorch_ops` work on GPU

1

This PR adds a small modification to the `NumpyOpWrapper` that is used in `pytorch_ops`, making the torch functions work with CUDA tensors. This should fix the following two issues: -...

Using --seed=0 will lead to non-deterministic behavior when using utils.setup()

2

I am concerned about the following line: https://github.com/google-research/kubric/blob/12d5aa4297e3de9cbfbb7bf9978366efe191e369/kubric/utils.py#L109 If we have `flags.seed=0`, `seed` will be set to `np.random.randint(0, 2147483647)`. This is obviously not the behavior that anybody would expect, as...

feat: implemented `RepeatAction` wrapper

1

# Description Implements feature request #652. This still needs a unit test. ## Type of change Please delete options that are not relevant. - [x] New feature (non-breaking change which...

Transition to Gymnasium / Future Compatibility with Tianshou

2

The maintainers of OpenAI Gym recently forked Gym to [Gymnasium](https://github.com/Farama-Foundation/Gymnasium), which is where all future maintenance will occur. More information can be found in the [announcement blog post](https://farama.org/Announcing-The-Farama-Foundation) of the...