C# Класс GoodAI.ToyWorld.Language.Vocabulary

Words and word vectors.
Показать файл Открыть проект Примеры использования класса

Открытые свойства

Свойство Тип Описание
VectorType WordVectorType

Открытые методы

Метод Описание
Add ( string label ) : void

Adds a labeled vector to the vocabulary, assigning a default (random) value to the vector.

Add ( string label, float vector ) : void

Adds a labeled vector to the vocabulary.

FindNearestNeighbors ( LabeledVector vector, int neighborhoodSize ) : LabeledVector>>.List

Finds the vectors most similar to the input vector. If the input is the zero vector, an empty list is returned.

FindNearestNeighbors ( float vector, int neighborhoodSize ) : LabeledVector>>.List

Finds the vectors most similar to the input vector. If the input is the zero vector, an empty list is returned.

IsZero ( float vector ) : bool

Returns true if the argument is the zero vector.

Read ( StreamReader vocabularyReader ) : void

Loads the vocabulary from text using a StreamReader

ReadTextFile ( string path ) : void

Loads the vocabulary from a file in text format (e.g., word2vec)

VectorFromLabel ( string label ) : float[]

Retrieves the vector that corresponds to a label.

Vocabulary ( int vectorDimensions, WordVectorType vectorType = WordVectorType.Random ) : System

Constructs the Vocabulary.

Приватные методы

Метод Описание
GetOOVVector ( string label ) : float[]

Returns a new vector for an out-of-vocabulary word.

MakeNewVector ( ) : float[]

Creates a new vector.

MakeOneOfNVector ( ) : float[]

Creates a new one-of-N vector.

MakeRandomVector ( ) : float[]

Creates a random vector uniformly distributed over the vector space. Components are in [0, 1[.

ReadFileHeader ( StreamReader vocabularyReader ) : void

Reads the first part of a text file. The header consists of a single line with one or two fields: [number_of_vectors] number_of_dimensions The word2vec text format uses both fields.

ReadLabeledVectors ( StreamReader vocabularyReader ) : void

Reads labeled vectors from a text file.

ReadLineFields ( StreamReader vocabularyReader ) : string[]

Reads a line and splits it on whitespace into fields.

Описание методов

Add() публичный Метод

Adds a labeled vector to the vocabulary, assigning a default (random) value to the vector.
public Add ( string label ) : void
label string The label
Результат void

Add() публичный Метод

Adds a labeled vector to the vocabulary.
public Add ( string label, float vector ) : void
label string The label
vector float The vector
Результат void

FindNearestNeighbors() публичный Метод

Finds the vectors most similar to the input vector. If the input is the zero vector, an empty list is returned.
public FindNearestNeighbors ( LabeledVector vector, int neighborhoodSize ) : LabeledVector>>.List
vector LabeledVector The input vector
neighborhoodSize int The number of neighbors to retrieve
Результат LabeledVector>>.List

FindNearestNeighbors() публичный Метод

Finds the vectors most similar to the input vector. If the input is the zero vector, an empty list is returned.
public FindNearestNeighbors ( float vector, int neighborhoodSize ) : LabeledVector>>.List
vector float The input vector
neighborhoodSize int The number of neighbors to retrieve
Результат LabeledVector>>.List

IsZero() публичный статический Метод

Returns true if the argument is the zero vector.
public static IsZero ( float vector ) : bool
vector float The input vector
Результат bool

Read() публичный Метод

Loads the vocabulary from text using a StreamReader
public Read ( StreamReader vocabularyReader ) : void
vocabularyReader System.IO.StreamReader The reader
Результат void

ReadTextFile() публичный Метод

Loads the vocabulary from a file in text format (e.g., word2vec)
public ReadTextFile ( string path ) : void
path string The path to the vocabulary text file
Результат void

VectorFromLabel() публичный Метод

Retrieves the vector that corresponds to a label.
public VectorFromLabel ( string label ) : float[]
label string The label to look up
Результат float[]

Vocabulary() публичный Метод

Constructs the Vocabulary.
public Vocabulary ( int vectorDimensions, WordVectorType vectorType = WordVectorType.Random ) : System
vectorDimensions int The number of word vector dimensions
vectorType WordVectorType The vector type used
Результат System

Описание свойств

VectorType публичное свойство

The vector type used (see WordVectorType)
public WordVectorType VectorType
Результат WordVectorType