https://ourbigbook.com/cirosantilli/farama-gymnasium I played a bit with Gymnasium https://github.com/Farama-Foundation/Gymnasium , a reinforcement learning environment forked from OpenAI Gym and maintained by the Farama Foundation. I understood the basic APIs and was glad to see that things just worked on my Ubuntu 24.10. Maybe one day I'll try my hand at actually implementing an agent. Though a more complicated and realistic robot task benchmark like Habitat or BEHAVIOUR seem more appealing. A quick video with manual control:
GIF