[Feature Request] Have a render() function exposed for Gym environments to see agent performance

[bmcnns]

Motivation

When I'm training a RL model cumulative reward isn't always a true indicator of model performance. It's helpful for me to be able to see how an agent is interacting with the environment to ensure that the training performance is matching my expectation.

Solution

Ideally, I'd like to have a method of visualizing the environments. In particular the Gym, ALE, and Mujoco environments.
Individually, these environments typically expose a render() function of RGB output. Are these methods exposed in the envpool implementation?

Alternatives

None come to mind.

Additional context

Checklist

• [ X ] I have checked that there is no similar issue in the repo (required)