ml

Documentation

Overview

Package ml provides some fundamental machine learning primitives.

Package ml provides some fundamental machine learning primitives.

Index

• func ConvertToFloat64Slice(slice interface{}) ([]float64, error)
• func CreateNewTensor(pft *pb.FlatTensor) (*tensor.Dense, error)
• func FormatClassificationOutputs(outNameMap *sync.Map, outMap Tensors, labels []string) (classification.Classifications, error)
• func TensorNames(t Tensors) []string
• type Tensors
  • func ProtoToTensors(pbft *pb.FlatTensors) (Tensors, error)

Functions

func ConvertToFloat64Slice

func ConvertToFloat64Slice(slice interface{}) ([]float64, error)

ConvertToFloat64Slice converts any numbers or slice of numbers into a float64 slice.

func CreateNewTensor

func CreateNewTensor(pft *pb.FlatTensor) (*tensor.Dense, error)

CreateNewTensor turns a proto FlatTensor into a *tensor.Dense.

func FormatClassificationOutputs

func FormatClassificationOutputs(
	outNameMap *sync.Map, outMap Tensors, labels []string,
) (classification.Classifications, error)

FormatClassificationOutputs formats the output tensors from a model into classifications.

func TensorNames

func TensorNames(t Tensors) []string

TensorNames returns all the names of the tensors.

Types

type Tensors

type Tensors map[string]*tensor.Dense

Tensors are a data structure to hold the input and output map of tensors that will fed into a model, or come from the result of a model.

func ProtoToTensors

func ProtoToTensors(pbft *pb.FlatTensors) (Tensors, error)

ProtoToTensors takes pb.FlatTensors and turns it into a Tensors map.