An excellent and comprehensive benchmark. But, the tasks list is desired.
Jinyi6 opened this issue · 1 comments
Jinyi6 commented
I appreciate such a wonderful benchmark. However, I think that the tasks list is desired.
such as:
https://github.com/openai/safety-gym#benchmark-suite
https://github.com/openai/gym/wiki/Table-of-environments
Thanks~
JacopoPan commented
Thanks @Jinyi6 ,
that's a good point: we are currently using a mix of the Gym-style registration system and our own YAML-based configuration system (which has defaults plus overrides, see safe_control_gym/envs/__init __.py).
I will come up with as simple a summary of the configurations I can once we have all the experiments from this paper online.