output

Constraints on the output responses of a model

Classes

Blank_Constraint(target)	Dummy constraint function that proposes all samples as feasible.
L1_Constraint(target[, offset])	Calculates the L1 (absolute-value penalized) constraint
Output_Constraint(target)	Output constraint for a given set of targets.

class obsidian.constraints.output.Blank_Constraint(target: Target | list[Target])

Bases: Output_Constraint

Dummy constraint function that proposes all samples as feasible.

forward(scale: bool = True) → Callable

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class obsidian.constraints.output.L1_Constraint(target: Target | list[Target], offset: int | float = 1)

Bases: Output_Constraint

Calculates the L1 (absolute-value penalized) constraint

forward(scale: bool = True) → Callable

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class obsidian.constraints.output.Output_Constraint(target: Target | list[Target])

Bases: Constraint

Output constraint for a given set of targets.

Must return a callable function that computes feasibility, where

negative values imply feasible space.

Note: Saving and loading input constraints is managed by Campaign