Optimization¶
The Gauss-Newton (GN) algorithm solving non-linear least squares problems. |
|
alias of |
|
The Levenberg-Marquardt (LM) algorithm solving non-linear least squares problems. |
|
alias of |
|
The batched linear solver with pseudo inversion. |
|
The batched linear solver with fast pseudo inversion. |
|
The batched linear solver with Cholesky decomposition. |
|
Constant damping strategy used in the Levenberg-Marquardt (LM) algorithm. |
|
Adaptive damping strategy used in the Levenberg-Marquardt (LM) algorithm. |
|
The trust region (TR) algorithm used in the Levenberg-Marquardt (LM) algorithm. |
|
The robust Scale kernel cost function. |
|
The robust Huber kernel cost function. |
|
The robust pseudo Huber kernel cost function. |
|
The robust SoftLOne kernel cost function. |
|
The robust Cauchy kernel cost function. |
|
The robust Arctan kernel cost function. |
|
The robust Tolerant kernel cost function. |
|
Faster yet stable version of Triggs correction of model residual and Jacobian. |
|
The Triggs correction of model residual and Jacobian. |
|
A scheduler to stop an optimizer when no relative loss 'decreasing' is seen for a 'patience' number of steps. |
|
Compute the model Jacobian with respect to the model parameters. |