Question about suppressing errors in `model.py`

[tabedzki]

In model.py there is a with np.errstate() that suppresses warnings about dividing by 0 except for the instance when we are doing 0/0 since by IEEE standards that is defined as a separate type of error.

Should we update this np.errstate() to account for both or should we not suppress the invalid state?

Thank you in advance