optimization problem with multple target variables

[mhayoz]

I have an optimization problem with two target variables in the following form:

$y_1, y_2 = \argmin_{y_1, y_2} f(x,y_1, y_2)$

where $y_1$ and $y_2$ do not have the same shape. Is it possible to use your pytorch implementation of DDN for this?

Thank you in advance for your answer.

[sgould]

The best way to handle this is to vectorize both y_1 and y_2 and concatenate them into a single variable for the purposes of differentiation. Within your optimisation procedure (and outside of the node) you can separate them again into different variables.