Crate argmin_core

Argmin Optimizaton toolbox core

This crate contains the core functionality of argmin. If you just want to run an optimization method, this is not what you are looking for. However, if you want to implement your own solver based on the argmin architecture, you should find all necessary tools here.

TODOs:

• Provide an example of how to implement a solver

Modules

finitediff	Finite Differentiation
macros	Macros
modcholesky	Modified Cholesky decompositions Modified Cholesky decompositions

Macros

check_param	Release an T from an Option<T> if it is not None. If it is None, return an ArgminError with a message that needs to be provided.
make_kv	Creates an ArgminKV at compile time in order to avoid pushing to the kv vector.

Structs

ArgminBase	Storage for data needed by most solvers
ArgminCheckpoint
ArgminIterData	The datastructure which is returned by the next_iter method of the ArgminIter trait.
ArgminKV	A simple key-value storage
ArgminLogger	Container for ArgminLoggers
ArgminResult	Return struct for all solvers.
ArgminSlogLogger	A logger based on slog
ArgminWriter
Error	The Error type, which can contain any failure.
MinimalNoOperator
NoOperator
WriteToFile

Enums

ArgminError
CheckpointMode
TerminationReason	Indicates why the optimization algorithm stopped

Traits

ArgminAdd	Add a T to self
ArgminDiv	(Pointwise) Divide a T by self
ArgminDot	Dot/scalar product of T and self
ArgminEye
ArgminInv	Compute the inverse (T) of self
ArgminIter	Main part of every solver: next_iter computes one iteration of the algorithm and init is executed before these iterations. The init method comes with a default implementation which corresponds to doing nothing.
ArgminLineSearch	Defines a common interface to line search methods. Requires that ArgminSolver is implemented for the line search method as well.
ArgminLog	Defince the interface every logger needs to expose
ArgminMul	(Pointwise) Multiply a T with self
ArgminNLCGBetaUpdate	Every method for the update of beta needs to implement this trait.
ArgminNorm	Compute the l2-norm (U) of self
ArgminOp	This trait needs to be implemented for every operator/cost function.
ArgminScaledAdd	Add a T scaled by an U to self
ArgminScaledSub	Subtract a T scaled by an U from self
ArgminSolver	Defines the interface to a solver. Usually, there is no need to implement this manually, use the argmin_derive crate instead.
ArgminSub	Subtract a T from self
ArgminTranspose
ArgminTrustRegion	Defines a common interface to methods which calculate approximate steps for trust region methods. Requires that ArgminSolver is implemented as well.
ArgminWeightedDot	Dot/scalar product of T and self weighted by W (p^TWv)
ArgminWrite	Every writer (which is something that writes parameter vectors somewhere after each iteration) needs to implement this.
ArgminZero	Return param vector of all zeros (for now, this is a hack. It should be done better)

Functions

load_checkpoint