Add examples for RL

[luigiberducci]

Hi,
What about adding single-file implementations for training agents in the gym?

I can think of stable baselines for single agent training. Any idea for multi-agent? @hzheng40

[naman-gupta99]

I would love to have this as I am struggling to get into RL and may be this will help me get some guidance on how to build some agents.