ML IERR: Difference between revisions

Revision as of 11:40, 25 January 2023

Default: ML_IERR	= 0	if ML_ISTART=2
	= 1	otherwise

Description: Calculation and output frequency of Bayesian error estimate.

Mind: This tag is only available in the development version of VASP.

This tag sets the distance in units of molecular-dynamics steps at which the energies, forces, stresses, etc. are written to the corresponding files (OUTCAR, OSZICAR, ML_LOGFILE, etc.). If learning is activated (ML_ISTART=0,1 or 3) the only allowed option is that Bayesian error estimation is "on" at every molecular-dynamics step (ML_IERR=1 is required).

Only for ML_ISTART=2 using the full algorithm the interval can be freely chosen via ML_IERR. No error estimation is available for the fast version (force field trained with ML_LFAST). For ML_IERR=0 the Bayesian error is never calculated (and also no corresponding log file line is written out). This is the default since for ML_ISTART=2 the Bayesian error estimation takes up a significant amount of the total calculation time.

Warning: ML_IERR>0 can never be used with the fast execution mode (force field trained with ML_LFAST=.TRUE.).

@@ Line 6: / Line 6: @@
 ----
-This tag sets the distance in units of molecular-dynamics steps at which the Bayesian error estimates are calculated and written to the {{TAG|ML_LOGFILE}}. That is, in every {{TAG|ML_IERR}} steps an entry corresponding to the keyword <code>BEE</code> and/or <code>BEEF</code> is written. If learning is activated ({{TAG|ML_ISTART}}=0,1 or 3) Bayesian error estimation must be "on" at every molecular-dynamics step ({{TAG|ML_IERR}}=1 is required).
+This tag sets the distance in units of molecular-dynamics steps at which the energies, forces, stresses, etc. are written to the corresponding files ({{TAG|OUTCAR}}, {{TAG|OSZICAR}}, {{TAG|ML_LOGFILE}}, etc.). If learning is activated ({{TAG|ML_ISTART}}=0,1 or 3) the only allowed option is that Bayesian error estimation is "on" at every molecular-dynamics step ({{TAG|ML_IERR}}=1 is required).
-Only for {{TAG|ML_ISTART}}=2 the interval can be freely chosen via {{TAG|ML_IERR}}. For {{TAG|ML_IERR}}=0 the Bayesian error is never calculated (and also no corresponding log file line is written out). This is the default since for {{TAG|ML_ISTART}}=2 the Bayesian error estimation takes up a significant amount of the total calculation time.
+Only for {{TAG|ML_ISTART}}=2 using the full algorithm the interval can be freely chosen via {{TAG|ML_IERR}}. No error estimation is available for the fast version (force field trained with {{TAG|ML_LFAST}}). For {{TAG|ML_IERR}}=0 the Bayesian error is never calculated (and also no corresponding log file line is written out). This is the default since for {{TAG|ML_ISTART}}=2 the Bayesian error estimation takes up a significant amount of the total calculation time.
-{{NB|warning|{{TAG|ML_IERR}}>0 can be also used with the fast execution mode {{TAG|ML_LFAST}}{{=}}.TRUE., but for each step executing the Bayesian error estimation {{VASP}} switches to the slower code branch. Hence, choosing a small interval for {{TAG|ML_IERR}} will result in a considerable slowdown of the code.}}
+{{NB|warning|{{TAG|ML_IERR}}>0 can never be used with the fast execution mode (force field trained with {{TAG|ML_LFAST}}{{=}}.TRUE.).}}
 == Related tags and articles ==