ai.djl

Interface Application.NLP

Enclosing class:

Application

public static interface Application.NLP

The common set of applications for natural language processing (text data).

Field Summary

Fields

Modifier and Type	Field and Description
static Application	ANY Any NLP application, including those in Application.NLP.
static Application	MACHINE_TRANSLATION An application that translates text from one language to another.
static Application	MULTIPLE_CHOICE An application to represent a multiple choice question.
static Application	QUESTION_ANSWER An application that a reference document and a question about the document and returns text answering the question.
static Application	SENTIMENT_ANALYSIS An application that classifies text into positive or negative, an specific case of TEXT_CLASSIFICATION.
static Application	TEXT_CLASSIFICATION An application that classifies text data.
static Application	TEXT_EMBEDDING An application that takes text and returns a feature vector that represents the text.
static Application	WORD_EMBEDDING An application that takes a word and returns a feature vector that represents the word.

Field Detail

ANY

static final Application ANY

Any NLP application, including those in Application.NLP.

QUESTION_ANSWER

static final Application QUESTION_ANSWER

An application that a reference document and a question about the document and returns text answering the question.

The typical signature is Model<QAInput, String>.

TEXT_CLASSIFICATION

static final Application TEXT_CLASSIFICATION

An application that classifies text data.

The typical signature is Model<String, Classifications>.

SENTIMENT_ANALYSIS

static final Application SENTIMENT_ANALYSIS

An application that classifies text into positive or negative, an specific case of TEXT_CLASSIFICATION.

WORD_EMBEDDING

static final Application WORD_EMBEDDING

An application that takes a word and returns a feature vector that represents the word.

The most representative signature is Model<String, NDArray>. However, many models will only embed a fixed Vocabulary of words. These words are usually given integer indices which may make the signature Model<String, NDArray> (or NDArray). The signatures may also use singleton NDLists instead of NDArray.

MACHINE_TRANSLATION

static final Application MACHINE_TRANSLATION

An application that translates text from one language to another.

The typical signature is Model<String, String>.

MULTIPLE_CHOICE

static final Application MULTIPLE_CHOICE

An application to represent a multiple choice question.

TEXT_EMBEDDING

static final Application TEXT_EMBEDDING

An application that takes text and returns a feature vector that represents the text.

The special case where the text consists of only a word is a WORD_EMBEDDING. The typical signature is Model<String, NDArray>.