Benchmark algorithms + wandb reports

[ffelten]

• [x] Define which algo should be tested on which environments
• [x] Create scripts for running the whole thing #36 #41

[PatrickYanZ]

it will be nice. Now, I rely on the logs from wandb and the python crawler grabs data and puts it into the pandas data frame. It is much easier to export step, total reward, discount reward to output files. wandb has some tools to implement this.

[ffelten]

see also #43

[ffelten]

See https://wandb.ai/openrlbenchmark/MORL-Baselines?workspace