TMVA::Experimental::RBDT< Backend > Class Template Reference

template<typename Backend = BranchlessJittedForest<float>>
class TMVA::Experimental::RBDT< Backend >

Fast boosted decision tree inference.

Definition at line 36 of file RBDT.hxx.

Public Types
using	Backend_t = Backend
using	Value_t = typename Backend::Value_t

Public Member Functions
	RBDT (const std::string &key, const std::string &filename)
	Construct backends from model in ROOT file.
RTensor< Value_t >	Compute (const RTensor< Value_t > &x)
	Compute model prediction on input RTensor.
std::vector< Value_t >	Compute (const std::vector< Value_t > &x)
	Compute model prediction on a single event.
template<typename Vector >
Vector	Compute (const Vector &x)
	Compute model prediction on a single event.

Private Attributes
std::vector< Backend_t >	fBackends
bool	fNormalizeOutputs
int	fNumOutputs

#include <TMVA/RBDT.hxx>

Member Typedef Documentation

Backend_t

template<typename Backend = BranchlessJittedForest<float>>

using TMVA::Experimental::RBDT< Backend >::Backend_t = Backend

Definition at line 39 of file RBDT.hxx.

Value_t

template<typename Backend = BranchlessJittedForest<float>>

using TMVA::Experimental::RBDT< Backend >::Value_t = typename Backend::Value_t

Definition at line 38 of file RBDT.hxx.

Constructor & Destructor Documentation

RBDT()

template<typename Backend = BranchlessJittedForest<float>>

TMVA::Experimental::RBDT< Backend >::RBDT ( const std::string &  key, const std::string &  filename  )

inline

Construct backends from model in ROOT file.

Definition at line 48 of file RBDT.hxx.

Member Function Documentation

Compute() [1/3]

template<typename Backend = BranchlessJittedForest<float>>

RTensor< Value_t > TMVA::Experimental::RBDT< Backend >::Compute ( const RTensor< Value_t > &  x )

inline

Compute model prediction on input RTensor.

Definition at line 99 of file RBDT.hxx.

Compute() [2/3]

template<typename Backend = BranchlessJittedForest<float>>

std::vector< Value_t > TMVA::Experimental::RBDT< Backend >::Compute ( const std::vector< Value_t > &  x )

inline

Compute model prediction on a single event.

Definition at line 96 of file RBDT.hxx.

Compute() [3/3]

template<typename Backend = BranchlessJittedForest<float>>

template<typename Vector >

Vector TMVA::Experimental::RBDT< Backend >::Compute ( const Vector &  x )

inline

Compute model prediction on a single event.

The method is intended to be used with std::vectors-like containers, for example RVecs.

Definition at line 79 of file RBDT.hxx.

Member Data Documentation

fBackends

template<typename Backend = BranchlessJittedForest<float>>

std::vector<Backend_t> TMVA::Experimental::RBDT< Backend >::fBackends

private

Definition at line 44 of file RBDT.hxx.

fNormalizeOutputs

template<typename Backend = BranchlessJittedForest<float>>

bool TMVA::Experimental::RBDT< Backend >::fNormalizeOutputs

private

Definition at line 43 of file RBDT.hxx.

fNumOutputs

template<typename Backend = BranchlessJittedForest<float>>

int TMVA::Experimental::RBDT< Backend >::fNumOutputs

private

Definition at line 42 of file RBDT.hxx.