SFCustomLanguageModelData

identifier

get identifier(): string

final let identifier: String

supportedPhonemes

static supportedPhonemes(locale: _Foundation_Locale): Array<string>

static func supportedPhonemes(locale: Locale) -> [String]

List the supported subset of X-SAMPA pronunciations supported by this locale for the Speech framework. SFCustomLanguageModelData accepts custom pronunciations whose phonetic representations are spelled out using an alphabet of pronunciations called the Extended Speech Assessment Methods Public Alphabet, or X-SAMPA. X-SAMPA consists of ASCII characters that, individually or in combination, represents sounds made by people when speaking, and is preferred in this context over the International Phonetic Alphabet (IPA) for ease of typing. Each locale supports only a subset of all possible sounds represented by X-SAMPA symbols, and this method exists to allow developers to query for the set of sounds that can be used for a given locale's custom pronunciations facility. - Parameters: - locale: the region and language whose supported pronunciations are being queried

export

export(path: string): Promise<void>

func export(to path: URL) async throws

Export the accumulated data to a file. The file produced by this method can be provided to `SFSpeechLanguageModel.prepareCustomLanguageModel` to produce language model and vocabulary files that are then ready to be used in conjunction with the `SFSpeechRecognizer`. - Parameters: - path: a URL where the exported data will be saved. - Throws: Errors related to creating directories and files, and deleting files

insert

insertPhraseCount(phraseCount: SFCustomLanguageModelData_PhraseCount): void

func insert(phraseCount: SFCustomLanguageModelData.PhraseCount)

Add a sample to the body of training data. This class accumulates text data that will later be used to train a language model, which can be provided to an `SFSpeechRecognizer` to improve performance on certain phrases. - Parameters: - phraseCount: A sample of text on which to train your custom language model

insert

insertPhraseCountGenerator(phraseCountGenerator: SFCustomLanguageModelData_PhraseCountGenerator): void

func insert(phraseCountGenerator: SFCustomLanguageModelData.PhraseCountGenerator)

Add a stream of samples to the body of training data. This class accumulates text data that will later be used to train a language model, which can be provided to an `SFSpeechRecognizer` to improve performance on certain phrases. - Parameters: - phraseCountGenerator: A generator of phrase counts

insert

insertTerm(term: SFCustomLanguageModelData_CustomPronunciation): void

func insert(term: SFCustomLanguageModelData.CustomPronunciation)

Add a custom term to the vocabulary. This class accumulates vocabulary data (in the form of tokens paired with X-SAMPA representations of the spoken forms of those tokens) which will later be processed and then provided to an `SFSpeechRecognizer`, to enable it to recognize words that are typically out-of-vocabulary. - Parameters: - term: A token, paired with an X-SAMPA representation of the token's pronunciation

locale

get locale(): _Foundation_Locale

final let locale: Locale

init

static init(locale: _Foundation_Locale, identifier: string, version: string): SFCustomLanguageModelData

init(locale: Locale, identifier: String, version: String)

Constructs an empty data container. The `SFCustomLanguageModelData` class accumulates language model training and custom vocabulary data, both associated with a specified locale. This initializer creates an object that initially holds no data. - Parameters: - locale: the region and language of the training data (must match with the locale used to construct the `SFSpeechRecognizer` later) - identifier: used to uniquely identify the resulting language model on the device where it will be processed - version: used to distinguish different versions of the language model on the device where it will be processed

version

get version(): string

final let version: String