treelite::gtil Namespace Reference

Classes
struct	Configuration
	Configuration class. More...

Enumerations
enum class	PredictKind : std::int8_t { kPredictDefault = 0 , kPredictRaw = 1 , kPredictLeafID = 2 , kPredictPerTree = 3 }
	Prediction type. More...

Functions
template<typename InputT >
void	Predict (Model const &model, InputT const *input, std::uint64_t num_row, InputT *output, Configuration const &config)
	Predict with dense data. More...
template<typename InputT >
void	PredictSparse (Model const &model, InputT const *data, std::uint64_t const *col_ind, std::uint64_t const *row_ptr, std::uint64_t num_row, InputT *output, Configuration const &config)
	Predict with sparse data with CSR (compressed sparse row) layout. More...
std::vector< std::uint64_t >	GetOutputShape (Model const &model, std::uint64_t num_row, Configuration const &config)
	Given a data matrix, query the necessary shape of array to hold predictions for all data points. More...
template void	Predict< float > (Model const &, float const *, std::uint64_t, float *, Configuration const &)
template void	Predict< double > (Model const &, double const *, std::uint64_t, double *, Configuration const &)
template void	PredictSparse< float > (Model const &, float const *, std::uint64_t const *, std::uint64_t const *, std::uint64_t, float *, Configuration const &)
template void	PredictSparse< double > (Model const &, double const *, std::uint64_t const *, std::uint64_t const *, std::uint64_t, double *, Configuration const &)

Enumeration Type Documentation

PredictKind

enum treelite::gtil::PredictKind : std::int8_t

strong

Prediction type.

Enumerator
kPredictDefault	Usual prediction method: sum over trees and apply post-processing. Expected output dimensions: (num_row, num_target, max_num_class)
kPredictRaw	Sum over trees, but don't apply post-processing; get raw margin scores instead. Expected output dimensions: (num_row, num_target, max_num_class)
kPredictLeafID	Output one (integer) leaf ID per tree. Expected output dimensions: (num_row, num_tree)
kPredictPerTree	Output one or more margin scores per tree. Expected output dimensions: (num_row, num_tree, leaf_vector_shape[0] * leaf_vector_shape[1])

Function Documentation

GetOutputShape()

std::vector<std::uint64_t> treelite::gtil::GetOutputShape	(	Model const &	model,
		std::uint64_t	num_row,
		Configuration const &	config
	)

Given a data matrix, query the necessary shape of array to hold predictions for all data points.

Parameters

model	Treelite Model object
num_row	Number of rows in the input
config	Configuration of GTIL predictor. Set this by calling TreeliteGTILParseConfig.

Returns

Array shape

Predict()

template<typename InputT >

void treelite::gtil::Predict	(	Model const &	model,
		InputT const *	input,
		std::uint64_t	num_row,
		InputT *	output,
		Configuration const &	config
	)

Predict with dense data.

Parameters

model	Treelite Model object
input	The 2D data array, laid out in row-major layout
num_row	Number of rows in the data matrix.
output	Pointer to buffer to store the output. Call GetOutputShape to get the amount of buffer you should allocate for this parameter.
config	Configuration of GTIL predictor

Predict< double >()

template void treelite::gtil::Predict< double >	(	Model const &	,
		double const *	,
		std::uint64_t	,
		double *	,
		Configuration const &
	)

Predict< float >()

template void treelite::gtil::Predict< float >	(	Model const &	,
		float const *	,
		std::uint64_t	,
		float *	,
		Configuration const &
	)

PredictSparse()

template<typename InputT >

void treelite::gtil::PredictSparse	(	Model const &	model,
		InputT const *	data,
		std::uint64_t const *	col_ind,
		std::uint64_t const *	row_ptr,
		std::uint64_t	num_row,
		InputT *	output,
		Configuration const &	config
	)

Predict with sparse data with CSR (compressed sparse row) layout.

In the CSR layout, data[row_ptr[i]:row_ptr[i+1]] store the nonzero entries of row i, and col_ind[row_ptr[i]:row_ptr[i+1]] stores the corresponding column indices.

Parameters

model	Treelite Model object
data	Nonzero elements in the data matrix
col_ind	Feature indices. col_ind[i] indicates the feature index associated with data[i].
row_ptr	Pointer to row headers. Length is [num_row] + 1.
num_row	Number of rows in the data matrix.
output	Pointer to buffer to store the output. Call GetOutputShape to get the amount of buffer you should allocate for this parameter.
config	Configuration of GTIL predictor

PredictSparse< double >()

template void treelite::gtil::PredictSparse< double >	(	Model const &	,
		double const *	,
		std::uint64_t const *	,
		std::uint64_t const *	,
		std::uint64_t	,
		double *	,
		Configuration const &
	)

PredictSparse< float >()

template void treelite::gtil::PredictSparse< float >	(	Model const &	,
		float const *	,
		std::uint64_t const *	,
		std::uint64_t const *	,
		std::uint64_t	,
		float *	,
		Configuration const &
	)