CatLearnPot Class Reference

#include <CatLearnPot.h>

Inheritance diagram for CatLearnPot:

Public Member Functions
	CatLearnPot (const Parameters &a_params)
void	train_optimize (const MatrixXd &features, const MatrixXd &targets) override
void	force (long nAtoms, const double *positions, const int *atomicNrs, double *forces, double *energy, double *variance, const double *box) override
bool	isThreadSafe () const noexcept override
	Whether this potential's force() can be called from multiple threads on the SAME instance.
Public Member Functions inherited from eonc::SurrogatePotential
	SurrogatePotential (PotType a_ptype, const Parameters &a_params)
virtual	~SurrogatePotential ()=default
bool	isSurrogate () const noexcept override
	Whether this is a surrogate (GP) potential.
std::tuple< double, AtomMatrix, double >	get_ef_var (const AtomMatrix pos, const VectorXi atmnrs, const Matrix3d box)
Public Member Functions inherited from eonc::Potential
	Potential (PotType a_ptype)
	Potential (PotType a_ptype, const Parameters &)
	Potential (const Parameters &a_params)
virtual	~Potential ()
std::tuple< double, AtomMatrix >	get_ef (const AtomMatrix &pos, const VectorXi &atmnrs, const Matrix3d &box)
PotType	getType () const
virtual bool	requiresIsolatedMoleculeLayout () const noexcept
	True for molecular QM / non-PBC backends (NWChem socket, ASE ORCA/NWChem, …).
virtual bool	isSharedInstanceThreadSafe () const noexcept
	Conservative gate for sharing one Potential instance across threads.
virtual bool	needsPerImageInstance () const noexcept
	Whether NEB should create separate Potential instances per image for true parallel force evaluation.
virtual bool	supportsBatchEvaluation () const noexcept
	Whether this potential supports batched evaluation of N systems in a single call.
virtual void	forceBatch (long nSystems, long nAtoms, const double *const *positions, const int *const *atomicNrs, double *const *forces, double *energies, double *variances, const double *const *boxes)
	Evaluate forces for N systems in a single call.

Public Attributes
py::object	m_gpmod
MatrixXd	variance
Public Attributes inherited from eonc::Potential
std::atomic< size_t >	forceCallCounter

Additional Inherited Members
Protected Attributes inherited from eonc::Potential
PotType	ptype

Detailed Description

Definition at line 25 of file CatLearnPot.h.

Constructor & Destructor Documentation

CatLearnPot()

CatLearnPot::CatLearnPot

(

const Parameters &

a_params

)

Definition at line 16 of file CatLearnPot.cpp.

    : SurrogatePotential(PotType::CatLearn, a_params) {
  py::module_ sys = py::module_::import("sys");
  py::exec(
      std::format("sys.path.insert(0, {})", a_params.catlearn_options.path));
 
  py::module_ gp_module = py::module_::import(
      "catlearn.regression.gaussianprocess.calculator.mlmodel");
 
  // Import the required modules
  // GP Model
  this->m_gpmod = gp_module.attr("get_default_model")(
      "model"_a = a_params.catlearn_options.model);
};

Member Function Documentation

force()

void CatLearnPot::force ( long nAtoms, const double * positions, const int * atomicNrs, double * forces, double * energy, double * variance, const double * box )

overridevirtual

Implements eonc::Potential.

Definition at line 36 of file CatLearnPot.cpp.

                                                             {
  MatrixXd features =
      Eigen::Map<MatrixXd>(const_cast<double *>(positions), 1, nAtoms * 3);
  py::tuple ef_and_unc = (this->m_gpmod.attr("predict")(
      features, "get_variance"_a = true, "get_derivatives"_a = true));
  auto ef_dat = ef_and_unc[0].cast<MatrixXd>();
  auto vari = ef_and_unc[1].cast<MatrixXd>();
  auto gradients = ef_dat.block(0, 1, 1, nAtoms * 3);
  for (int idx = 0; idx < nAtoms; idx++) {
    forces[3 * idx] = gradients(0, 3 * idx) * -1;
    forces[3 * idx + 1] = gradients(0, 3 * idx + 1) * -1;
    forces[3 * idx + 2] = gradients(0, 3 * idx + 2) * -1;
  }
  *variance = vari(0, 0); // energy variance only
  *energy = ef_dat(0, 0);
  return;
}

isThreadSafe()

bool CatLearnPot::isThreadSafe ( ) const

inlinenodiscardoverridevirtualnoexcept

Whether this potential's force() can be called from multiple threads on the SAME instance.

Python-based potentials return false. Potentials with internal mutex (MetatomicPotential) return true but serialize internally – use needsPerImageInstance() to check if separate instances would enable true parallelism.

Reimplemented from eonc::Potential.

Definition at line 36 of file CatLearnPot.h.

{ return false; }

train_optimize()

void CatLearnPot::train_optimize ( const MatrixXd & features, const MatrixXd & targets )

overridevirtual

Implements eonc::SurrogatePotential.

Definition at line 31 of file CatLearnPot.cpp.

                                                                    {
  m_gpmod.attr("optimize")(features, targets, py::arg("retrain") = true);
  return;
}

Member Data Documentation

m_gpmod

py::object CatLearnPot::m_gpmod

Definition at line 38 of file CatLearnPot.h.

variance

MatrixXd CatLearnPot::variance

Definition at line 40 of file CatLearnPot.h.

---

The documentation for this class was generated from the following files:

• /home/runner/work/eOn/eOn/include/eon/potentials/CatLearnPot/CatLearnPot.h
• /home/runner/work/eOn/eOn/client/potentials/CatLearnPot/CatLearnPot.cpp