JavaScript (NodeJS / ElectronJS)

Model

class Model(aModelPath)

An object providing an interface to a trained DeepSpeech model.

Arguments

  • aModelPath (string) – The path to the frozen model graph.

Throws

on error

Model.beamWidth()

Get beam width value used by the model. If :js:func:Model.setBeamWidth was not called before, will return the default value loaded from the model file.

Returns

number – Beam width value used by the model.

Model.createStream()

Create a new streaming inference state. One can then call Stream.feedAudioContent() and Stream.finishStream() on the returned stream object.

Throws

on error

Returns

object – a Stream() object that represents the streaming state.

Model.disableExternalScorer()

Disable decoding using an external scorer.

Returns

number – Zero on success, non-zero on failure (invalid arguments).

Model.enableExternalScorer(aScorerPath)

Enable decoding using an external scorer.

Arguments

  • aScorerPath (string) – The path to the external scorer file.

Returns

number – Zero on success, non-zero on failure (invalid arguments).

Model.sampleRate()

Return the sample rate expected by the model.

Returns

number – Sample rate.

Model.setBeamWidth(The)

Set beam width value used by the model.

Arguments

  • The (number) – beam width used by the model. A larger beam width value generates better results at the cost of decoding time.

Returns

number – Zero on success, non-zero on failure.

Model.setScorerAlphaBeta(aLMAlpha, aLMBeta)

Set hyperparameters alpha and beta of the external scorer.

Arguments

  • aLMAlpha (float) – The alpha hyperparameter of the CTC decoder. Language Model weight.

  • aLMBeta (float) – The beta hyperparameter of the CTC decoder. Word insertion weight.

Returns

number – Zero on success, non-zero on failure (invalid arguments).

Model.stt(aBuffer)

Use the DeepSpeech model to perform Speech-To-Text.

Arguments

  • aBuffer (object) – A 16-bit, mono raw audio signal at the appropriate sample rate (matching what the model was trained on).

Returns

string – The STT result. Returns undefined on error.

Model.sttWithMetadata(aBuffer, aNumResults)

Use the DeepSpeech model to perform Speech-To-Text and output results including metadata.

Arguments

  • aBuffer (object) – A 16-bit, mono raw audio signal at the appropriate sample rate (matching what the model was trained on).

  • aNumResults (number) – Maximum number of candidate transcripts to return. Returned list might be smaller than this. Default value is 1 if not specified.

Returns

object – Metadata() object containing multiple candidate transcripts. Each transcript has per-token metadata including timing information. The user is responsible for freeing Metadata by calling FreeMetadata(). Returns undefined on error.

Stream

class Stream(nativeStream)

Provides an interface to a DeepSpeech stream. The constructor cannot be called directly, use Model.createStream().

Stream.feedAudioContent(aBuffer)

Feed audio samples to an ongoing streaming inference.

Arguments

  • aBuffer (buffer) – An array of 16-bit, mono raw audio samples at the appropriate sample rate (matching what the model was trained on).

Stream.finishStream()

Compute the final decoding of an ongoing streaming inference and return the result. Signals the end of an ongoing streaming inference.

Returns

string – The STT result. This method will free the stream, it must not be used after this method is called.

Stream.finishStreamWithMetadata(aNumResults)

Compute the final decoding of an ongoing streaming inference and return the results including metadata. Signals the end of an ongoing streaming inference.

Arguments

  • aNumResults (number) – Maximum number of candidate transcripts to return. Returned list might be smaller than this. Default value is 1 if not specified.

Returns

object – Outputs a Metadata() struct of individual letters along with their timing information. The user is responsible for freeing Metadata by calling FreeMetadata(). This method will free the stream, it must not be used after this method is called.

Stream.intermediateDecode()

Compute the intermediate decoding of an ongoing streaming inference.

Returns

string – The STT intermediate result.

Stream.intermediateDecodeWithMetadata(aNumResults)

Compute the intermediate decoding of an ongoing streaming inference, return results including metadata.

Arguments

  • aNumResults (number) – Maximum number of candidate transcripts to return. Returned list might be smaller than this. Default value is 1 if not specified.

Returns

object – Metadata() object containing multiple candidate transcripts. Each transcript has per-token metadata including timing information. The user is responsible for freeing Metadata by calling FreeMetadata(). Returns undefined on error.

Module exported methods

FreeModel(model)

Frees associated resources and destroys model object.

Arguments

  • model (object) – A model pointer returned by Model()

FreeStream(stream)

Destroy a streaming state without decoding the computed logits. This can be used if you no longer need the result of an ongoing streaming inference and don’t want to perform a costly decode operation.

Arguments
FreeMetadata(metadata)

Free memory allocated for metadata information.

Arguments

  • metadata (object) – Object containing metadata as returned by Model.sttWithMetadata() or Model.finishStreamWithMetadata()

Version()

Print version of this library and of the linked TensorFlow library on standard output.

Metadata

class Metadata()

An array of CandidateTranscript objects computed by the model.

Metadata.transcripts()

Array of transcripts

Returns

array – Array of CandidateTranscript() objects

CandidateTranscript

class CandidateTranscript()

A single transcript computed by the model, including a confidence value and the metadata for its constituent tokens.

CandidateTranscript.confidence()

Approximated confidence value for this transcription. This is roughly the sum of the acoustic model logit values for each timestep/token that contributed to the creation of this transcription.

Returns

float – Confidence value

CandidateTranscript.tokens()

Array of tokens

Returns

array – Array of TokenMetadata()

TokenMetadata

class TokenMetadata()

Stores text of an individual token, along with its timing information

TokenMetadata.start_time()

Position of the token in seconds

Returns

float – The position of the token

TokenMetadata.text()

The text corresponding to this token

Returns

string – The text generated

TokenMetadata.timestep()

Position of the token in units of 20ms

Returns

int – The position of the token