ML MHIS: Difference between revisions

From VASP Wiki
No edit summary
No edit summary
Line 4: Line 4:
----
----


This flag is only used if {{TAG|ML_ICRITERIA}}=''.TRUE.'' is set. The ML code stores {{TAG|ML_MHIS}} Bayesian errors from previous training steps. Specifically, immediately after a training step, the Bayesian error estimate for the forces are reevaluated for the just added structure, and the maximum error of the forces is stored in the history.  
This flag is only used if {{TAG|ML_ICRITERIA}} > 0. The ML code stores {{TAG|ML_MHIS}} Bayesian errors from previous training steps. Immediately after a training step, the Bayesian error estimate for the forces are reevaluated for the just added structure, and the maximum error of the forces is stored in the history.  


== Related Tags and Sections ==
== Related Tags and Sections ==

Revision as of 20:49, 11 October 2021

ML_MHIS = [integer]
Default: ML_MHIS = 10 

Description: This flag sets the number of estimated errors stored in memory to determine the threshold for the Bayesian error in the machine learning force field method.


This flag is only used if ML_ICRITERIA > 0. The ML code stores ML_MHIS Bayesian errors from previous training steps. Immediately after a training step, the Bayesian error estimate for the forces are reevaluated for the just added structure, and the maximum error of the forces is stored in the history.

Related Tags and Sections

ML_LMLFF, ML_ICRITERIA, ML_CTIFOR

Examples that use this tag