Allow the use of gymnasium >0.29.1

[traversaro]

I am looking into allowing to use lerobot with gymnasium 1.* . As several changes in several repos are necessary for this, I think it make sense to have an issue to track this.

• Fix gym-xarm to work with gymnasium >0.29.1 : Fix compatibility with gymnasium >= 1.0.0 gym-xarm#12
• Release gym-xarm with gymnasium >0.29.1 fixes
• Fix use of final_info in lerobot/scripts/eval.py, see Success rate computation for diffusion_policy pushT is broken since the release of the V-1.0.0 of the gymnasium #470 and Fix gymnasium version as pre-1.0.0 #471

[jadechoghari]

Hi @traversaro , we are currenlty wokring on upgrading all of our envs to gymnasium 1.xx along with mujoco >= 3.0 .
We will upade you here once this is done. Thank you!

[traversaro]

Hi @traversaro , we are currenlty wokring on upgrading all of our envs to gymnasium 1.xx along with mujoco >= 3.0 . We will upade you here once this is done. Thank you!

Great news, thanks!

[traversaro]

Thanks @jadechoghari for the hard work on this!

For reference the related PRs are probably:

Fix gym-xarm to work with gymnasium >0.29.1 : huggingface/gym-xarm#12
Release gym-xarm with gymnasium >0.29.1 fixes

The gym-xarm dependency was removed in #2088 .

Fix use of final_info in lerobot/scripts/eval.py, see #470 and #471

I could not find anything related to that, is there are any test that verify that there was no regression w.r.t. to #470 ?