Constraints¤
Constraints define objective terms for training or evaluation. They operate on domain functions and typically return a scalar loss term.
For the mathematical conventions used by sampled constraints (residual norms, reduction modes, measures, and filtering), see Guides → Constraints and objectives.