Continuous action space must have a finite lower and upper bound

Summary

If the variable ‘is_agent_action’ is set in the state config without an upper or lower limit, the state sets the upper and lower value to the default np.inf. The control instance of action_space from stables_baselines exspects finite lower and upper limits.

What is the expected correct behavior? What was your goal?

If the limit values are not defined in the state_config, the correct default values should be set.

Was there an error message?

File "C:\Users\xxx\Documents\gitProjects\dr-on-kea\.venv\Lib\site-packages\stable_baselines3\common\base_class.py", line 198, in __init__ assert np.all( ^^^^^^^ AssertionError: Continuous action space must have a finite lower and upper bound