You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This really complicates the code. I think the algorithm will be much easier if we by default initialize T_exp = zeros and have T_exp! = (u,p,t) -> nothing. The optimization as it is, is really only benefiting toy examples where, for example, T_exp (or T_lim) don't exist. If they're toy examples, adding these zeros (@. U_exp += dt * T_exp) is should not be an issue. Basically, I think we shouldn't be optimizing for toy examples, and write the code in a simpler way for the real world example, but also allow it to work for toy examples.
The text was updated successfully, but these errors were encountered:
One pattern that we're currently using is:
...
This really complicates the code. I think the algorithm will be much easier if we by default initialize
T_exp = zeros
and haveT_exp! = (u,p,t) -> nothing
. The optimization as it is, is really only benefiting toy examples where, for example,T_exp
(orT_lim
) don't exist. If they're toy examples, adding these zeros (@. U_exp += dt * T_exp
) is should not be an issue. Basically, I think we shouldn't be optimizing for toy examples, and write the code in a simpler way for the real world example, but also allow it to work for toy examples.The text was updated successfully, but these errors were encountered: