Closed timmyfaraday closed 3 years ago
It might be the case that additional bounds on all variables will help convergence, some input/feedback here especially for the non-expectation network would be helpful.
Currently, the bounds are too restrictive. Adding them results in another solution compared to the 'unbounded' case.
It might be the case that additional bounds on all variables will help convergence, some input/feedback here especially for the non-expectation network would be helpful.