model

package
v0.0.0-...-7d47eef Latest Latest
Warning

This package is not in the latest version of its module.

Go to latest
Published: Apr 12, 2023 License: Apache-2.0, MIT Imports: 2 Imported by: 0

Documentation

Overview

Package model defines the "Model" interface.

Index

Constants

This section is empty.

Variables

View Source
var RegisteredBuilders = make(map[string]func(header *model_pb.AbstractModel, dataspec *dataspec_pb.DataSpecification) Implementation)

RegisteredBuilders is the list of model builders, keyed by a unique `ModelKey` string per model type. Only register (change) this during the runtime initialization, in `init()` function. End users probably want to use `io.LoadModel()` to load models instead.

Functions

This section is empty.

Types

type Implementation

type Implementation interface {
	Model

	// LoadSpecific loads the model implementation specific data from a directory.
	// Users are not expected to call this method directly. Instead, models
	// should be loaded with "model, err := io.LoadModel(path)".
	LoadSpecific(modelPath string, prefix string) error
}

Implementation interface needs to be implemented by Models (mostly internal). Not needed by those only using a model.

type Model

type Model interface {

	// Registered name of the model.
	Name() string

	// Header of the model.
	Header() *model_pb.AbstractModel

	// Dataspec of the model.
	Dataspec() *dataspec_pb.DataSpecification
}

Model is a generic model interface.

Examples:

// Create a new GBT model model := gradientboostedtrees.Create(...)

// Load an existing model. model, err := io.Load("/path/to/model") fmt.Println("My model is a %v.", model.Name()) >> My model is a GRADIENT_BOOSTED_TREES.

Directories

Path Synopsis
Package canonical registers the "canonical" models.
Package canonical registers the "canonical" models.
Package decisiontree contains utilities to handle decision trees.
Package decisiontree contains utilities to handle decision trees.
io
Package io contains utilities to load/save decision trees
Package io contains utilities to load/save decision trees
io/blobsequence
Package blobsequence implement node reading from blob sequence files.
Package blobsequence implement node reading from blob sequence files.
io/canonical
Package canonical registers the "canonical" ways to read tree nodes.
Package canonical registers the "canonical" ways to read tree nodes.
proto
Package proto includes all proto definitions used in the golang package in one large package.
Package proto includes all proto definitions used in the golang package in one large package.
Package gradientboostedtrees defines the gradient boosted trees model.
Package gradientboostedtrees defines the gradient boosted trees model.
proto
Package proto includes all proto definitions used in the golang package in one large package.
Package proto includes all proto definitions used in the golang package in one large package.
io
Package io contains utilities to save and load models.
Package io contains utilities to save and load models.
canonical
Package canonical is an alias for the `model/io` package that also links all the canonical model support along.
Package canonical is an alias for the `model/io` package that also links all the canonical model support along.
Package proto includes all proto definitions used in the golang package in one large package.
Package proto includes all proto definitions used in the golang package in one large package.
Package randomforest defines the gradient boosted trees model.
Package randomforest defines the gradient boosted trees model.
proto
Package proto includes all proto definitions used in the golang package in one large package.
Package proto includes all proto definitions used in the golang package in one large package.

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL