What is the recommended technique for obtaining rendered videos of episodes?

[vrn-sn]

It looks like gymnasium.wrappers.Monitor is no longer supported, and the following RecordVideo wrapper doesn't seem to internally support the cost output that is returned when step is called:

from gymnasium.wrappers.record_video import RecordVideo

env = safety_gymnasium.make("SafetyRacecarGoal1-v0", render_mode="rgb_array")
env = RecordVideo(env, './videos')

I've also seen others render videos from a 3rd-person perspective, but I can only manage to get 1st-person perspective videos rendered. Does anyone know how to change this?

[muchvo]

Please refer to documentation for API and camera options on rendering the image, and follow this example file for saving images accordingly. For complete example, you can refer to here.