Describe the bug
I found that some names agruments in framework aren't consistent.
So for example:

class SupervisedRunner(Runner):
    """Runner for experiments with supervised model."""

    _experiment_fn: Callable = SupervisedExperiment

    def __init__(
        self,
        model: Model = None,
        device: Device = None,
        input_key: Any = "features", 
      

According to @rspadim, functions in entropy.py could use numpy array instead of strings, as it's better to numba.

Guide on how to implement this is available in the comments in PR #311 .

Self-explanatory.

🚀 Feature

360 sensor -> perspective view observation transformer

I would propose two observation transformers for this:

• Equirect -> perspective imgs (an observation transfomer that can extract multiple perspective imgs from an equirect input)
• A special case of this: Equirect -> Cubemap (allows policies trained on cubem