fix/gym_robotics

[m-stoll]

Use GoalEnv from gym_robotics (see openai/gym#2516)

[m-stoll]

Hmm, not as easy as we thought...
Maybe you'll figure out a solution easily.
Also this doesn't look right:
https://github.com/Farama-Foundation/gym-robotics/blob/main/setup.py#L25

[eleurent]

Hi,
Thank you for the suggestion!
I don't have permission to push on you PR branch, so I'll just merge this and try to fix it afterwards. I can see 3 ways to deal with this:

• install gym_robotics from their master, rather than pypi
• require gym version 0.21 (prior to the change), like it is done in StableBaselines3
• don't subclass GoalEnv anymore, but just implement the compute_rewards method.

[eleurent]

Solution 1 worked :)