gpr-opthyprms
Navigation: Documentation / Utilities / gpr-opthyprms
gpr-opthyprms: version 6.380.1d87074 (2021-07-26 16:36)
Name
gpr-opthyprms
Synopsis
gpr-opthyprms [OPTIONS] [--] accuname1 realm1 [accuname2 realm2 [..]] hyprmsname
Description
The program finds optimal GPR hyperparameters, which maximize logarithm of marginal likelihood. The optimization is performed by the L-BFGS optimizer employing either analytical or numerical gradients of logML w.r.t. hyperparameters.
Arguments
accuname1 Name of file containing the PMF accumulator.
realm1 Realm of interest.
hyprmsname Name of file where the optimized GPR hyperparameters are saved. If the name is '-' then the output will be written to the standard output.
realm1 Realm of interest.
hyprmsname Name of file where the optimized GPR hyperparameters are saved. If the name is '-' then the output will be written to the standard output.
Options
| -t, --target NAME Specify optimized targed, which can be either logml (log of marginal likelihood) or logpl (log of pseudo-likelihood from leave-one-out cross-validation (LOO-CV)). (Default: logml)
|
| -a, --lmethod NAME Linear algebra method for matrix inversion. Supported algorithms are: default, svd (SVD - singular value decomposition, divide and conquer driver), svd2 (SVD - singular value decomposition, simple driver), lu (LU factorization), and ll (LL - Cholesky factorization). (Default: default)
|
| -r, --rcond NUMBER Rank condition for SVD. Used value must be carefully tested. Calculation at computer precision is requested with -1 (not recommended). (Default: 1e-06)
|
| -s, --sigmaf2 NUMBER Variance of the reconstructed free energy surface (signal variance). (Default: 15)
|
| -c, --ncorr NUMBER Number of statistically correlated samples. (Default: 1)
|
| -w, --wfac SPEC Factors influencing widths of RBFs or square exponential kernels. The width is distance between the adjacent square exponential functions multiplied by this factors in the form WFac1[xWFac2x...]. The last value pads the rest. (Default: 3.0)
|
| --enablesigmaf2 Enable optimization of SigmaF2 hyperparameter.
|
| --enablencorr Enable optimization of NCorr hyperparameter.
|
| --enablewfac SPEC Enable optimization of Wfac hyperparamters. Flags are specified in the form WFac1Enabled[xWFac2Enabledx...] with F and T for disabled and enabled, respectively. The last value pads the rest. (Default: F)
|
| --numeric Use numeric gradients.
|
| --noptsteps NUMBER Maximum number of optimization steps. (Default: 100)
|
| --nresets NUMBER Maximum number of resets due to insufficient optimization progress. (Default: 3)
|
| --termeps NUMBER Termination criteria for L-BFGS optimizer (see L-BFGS code). (Default: 1e-05)
|
| --termval NUMBER Termination criteria for L-BFGS optimizer. Minimum change of optimized property. (Default: 1e-06)
|
| --nlbfgscorr NUMBER Number of corrections used in an L-BFGS update. (Default: 10)
|
| --test Compare analytical and numerical gradient for input hyperparameters.
|
| --cd5 Use 5-point stencil for numerical differentiation.
|
| --loadhyprms NAME Name of file with GPR hyperparameters.
|
| --stat Calculate detailed GPR status employing found hyperparameters.
|
| --sptype Determine type of stationary point.
|
| --kernel NAME GPR: Kernel type. Supported types: ardse (ARD squared exponential), ardmc52 (ARD Matern class 5/2), default(=ardse) (Default: default)
|
| --useinv Use matrix inversion pathway (for testing only).
|
| --calclogpl Calculate logPL for --stat if --target is not logpl.
|
| --globalmin SPEC position of global minimum provided as a single string in the form CV1xCV2x...xCVn (relevant for error determination), if not set the position is determined automatically.
|
| -v, --verbose Increase output verbosity.
|
| --version Output version information and exit.
|
| -h, --help Display this help and exit.
|