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?
In
model.py
there is awith np.errstate()
that suppresses warnings about dividing by 0 except for the instance when we are doing0/0
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