MBPO Implementation Example

[CAmpatzis]

🚀 Feature Request

Would it be possible to create an MBPO Jupyter notebook in a similar manner to that of PETS?

Motivation

I am currently conducting some research on the use of model-based reinforcement learning. While the MBPO "mbpo.train()" function is useful it does not fit all cases, especially concerning logging. It would incredibly useful if you created a Jupyter notebook showcasing how the MBPO training loop actually works.

The possibility of a PlaNet example would also be wonderful, however, I think MBPO would set a good baseline.

Thank you very much for your effort in developing this library, and consideration to my request.