SOMBRL: Configs for paper experiments

[jws-1]

Hi,

Thank you for open sourcing your implementations!

I was wondering if it's possible to share the configs used in the SOMBRL paper, specifically for the MBPO experiments? I notice that currently experiments/sombrl/launcher.py on the clean branch just has cartpole-swingup_sparse and quadruped-run configs, but I'd like to try and replicate the results on the other environments used in the paper.

Thanks!

Jared.