.NET Framework¶

DeepSpeech Class¶

class

Concrete implementation of DeepSpeechClient.Interfaces.IDeepSpeech.

Public Functions

DeepSpeechClient.DeepSpeech.DeepSpeech(string aModelPath)

Initializes a new instance of DeepSpeech class and creates a new acoustic model.

Parameters

aModelPath: The path to the frozen model graph.

Exceptions

ArgumentException: Thrown when the native binary failed to create the model.

unsafe uint DeepSpeechClient.DeepSpeech.GetModelBeamWidth()

Get beam width value used by the model. If SetModelBeamWidth was not called before, will return the default value loaded from the model file.

Return: Beam width value used by the model.

unsafe void DeepSpeechClient.DeepSpeech.SetModelBeamWidth(uint aBeamWidth)

Set beam width value used by the model.

Parameters

aBeamWidth: The beam width used by the decoder. A larger beam width value generates better results at the cost of decoding time.

Exceptions

ArgumentException: Thrown on failure.

unsafe void DeepSpeechClient.DeepSpeech.AddHotWord(string aWord, float aBoost)

Add a hot-word.

Parameters

aWord: Some word
aBoost: Some boost

Exceptions

ArgumentException: Thrown on failure.

unsafe void DeepSpeechClient.DeepSpeech.EraseHotWord(string aWord)

Erase entry for a hot-word.

Parameters

aWord: Some word

Exceptions

ArgumentException: Thrown on failure.

unsafe void DeepSpeechClient.DeepSpeech.ClearHotWords()

Clear all hot-words.

Exceptions

ArgumentException: Thrown on failure.

unsafe int DeepSpeechClient.DeepSpeech.GetModelSampleRate()

Return the sample rate expected by the model.

Return: Sample rate.

unsafe void DeepSpeechClient.DeepSpeech.Dispose(): Frees associated resources and destroys models objects.

unsafe void DeepSpeechClient.DeepSpeech.EnableExternalScorer(string aScorerPath)

Enable decoding using an external scorer.

Parameters

aScorerPath: The path to the external scorer file.

Exceptions

ArgumentException: Thrown when the native binary failed to enable decoding with an external scorer.
FileNotFoundException: Thrown when cannot find the scorer file.

unsafe void DeepSpeechClient.DeepSpeech.DisableExternalScorer()

Disable decoding using an external scorer.

Exceptions

ArgumentException: Thrown when an external scorer is not enabled.

unsafe void DeepSpeechClient.DeepSpeech.SetScorerAlphaBeta(float aAlpha, float aBeta)

Set hyperparameters alpha and beta of the external scorer.

Parameters

aAlpha: The alpha hyperparameter of the decoder. Language model weight.
aBeta: The beta hyperparameter of the decoder. Word insertion weight.

Exceptions

ArgumentException: Thrown when an external scorer is not enabled.

unsafe void DeepSpeechClient.DeepSpeech.FeedAudioContent(DeepSpeechStream stream, short [] aBuffer, uint aBufferSize)

Feeds audio samples to an ongoing streaming inference.

Parameters

stream: Instance of the stream to feed the data.
aBuffer: An array of 16-bit, mono raw audio samples at the appropriate sample rate (matching what the model was trained on).

unsafe string DeepSpeechClient.DeepSpeech.FinishStream(DeepSpeechStream stream)

Closes the ongoing streaming inference, returns the STT result over the whole audio signal.

Return

The STT result.

Parameters

stream: Instance of the stream to finish.

unsafe Metadata DeepSpeechClient.DeepSpeech.FinishStreamWithMetadata(DeepSpeechStream stream, uint aNumResults)

Closes the ongoing streaming inference, returns the STT result over the whole audio signal, including metadata.

Return

The extended metadata result.

Parameters

stream: Instance of the stream to finish.
aNumResults: Maximum number of candidate transcripts to return. Returned list might be smaller than this.

unsafe string DeepSpeechClient.DeepSpeech.IntermediateDecode(DeepSpeechStream stream)

Computes the intermediate decoding of an ongoing streaming inference.

Return

The STT intermediate result.

Parameters

stream: Instance of the stream to decode.

unsafe Metadata DeepSpeechClient.DeepSpeech.IntermediateDecodeWithMetadata(DeepSpeechStream stream, uint aNumResults)

Computes the intermediate decoding of an ongoing streaming inference, including metadata.

Return

The STT intermediate result.

Parameters

stream: Instance of the stream to decode.
aNumResults: Maximum number of candidate transcripts to return. Returned list might be smaller than this.

unsafe string DeepSpeechClient.DeepSpeech.Version(): Return version of this library. The returned version is a semantic version (SemVer 2.0.0).

unsafe DeepSpeechStream DeepSpeechClient.DeepSpeech.CreateStream(): Creates a new streaming inference state.

unsafe void DeepSpeechClient.DeepSpeech.FreeStream(DeepSpeechStream 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.

unsafe string DeepSpeechClient.DeepSpeech.SpeechToText(short [] aBuffer, uint aBufferSize)

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

Return

The STT result. Returns NULL on error.

Parameters

aBuffer: A 16-bit, mono raw audio signal at the appropriate sample rate (matching what the model was trained on).
aBufferSize: The number of samples in the audio signal.

unsafe Metadata DeepSpeechClient.DeepSpeech.SpeechToTextWithMetadata(short [] aBuffer, uint aBufferSize, uint aNumResults)

Use the DeepSpeech model to perform Speech-To-Text, return results including metadata.

Return

The extended metadata. Returns NULL on error.

Parameters

aBuffer: A 16-bit, mono raw audio signal at the appropriate sample rate (matching what the model was trained on).
aBufferSize: The number of samples in the audio signal.
aNumResults: Maximum number of candidate transcripts to return. Returned list might be smaller than this.

DeepSpeechStream Class¶

class DeepSpeechStream : public IDisposable¶

Wrapper of the pointer used for the decoding stream.

Public Functions

unsafe DeepSpeechClient.Models.DeepSpeechStream.DeepSpeechStream(IntPtr ** streamingStatePP)

Initializes a new instance of DeepSpeechStream.

Parameters

streamingStatePP: Native pointer of the native stream.

ErrorCodes¶

See also the main definition including descriptions for each error in Error codes.

enum DeepSpeechClient::Enums::ErrorCodes¶

Error codes from the native DeepSpeech binary.

Values:

DS_ERR_OK = 0x0000¶

DS_ERR_NO_MODEL = 0x1000¶

DS_ERR_INVALID_ALPHABET = 0x2000¶

DS_ERR_INVALID_SHAPE = 0x2001¶

DS_ERR_INVALID_SCORER = 0x2002¶

DS_ERR_MODEL_INCOMPATIBLE = 0x2003¶

DS_ERR_SCORER_NOT_ENABLED = 0x2004¶

DS_ERR_FAIL_INIT_MMAP = 0x3000¶

DS_ERR_FAIL_INIT_SESS = 0x3001¶

DS_ERR_FAIL_INTERPRETER = 0x3002¶

DS_ERR_FAIL_RUN_SESS = 0x3003¶

DS_ERR_FAIL_CREATE_STREAM = 0x3004¶

DS_ERR_FAIL_READ_PROTOBUF = 0x3005¶

DS_ERR_FAIL_CREATE_SESS = 0x3006¶

DS_ERR_FAIL_INSERT_HOTWORD = 0x3008¶

DS_ERR_FAIL_CLEAR_HOTWORD = 0x3009¶

DS_ERR_FAIL_ERASE_HOTWORD = 0x3010¶

Metadata¶

class Metadata¶

Stores the entire CTC output as an array of character metadata objects.

Property

property DeepSpeechClient::Models::Metadata::Transcripts: List of candidate transcripts.

CandidateTranscript¶

class CandidateTranscript¶

Stores the entire CTC output as an array of character metadata objects.

Property

property DeepSpeechClient::Models::CandidateTranscript::Confidence: Approximated confidence value for this transcription.

property DeepSpeechClient::Models::CandidateTranscript::Tokens: List of metada tokens containing text, timestep, and time offset.

TokenMetadata¶

class TokenMetadata¶

Stores each individual character, along with its timing information.

Public Members

string DeepSpeechClient.Models.TokenMetadata.Text: Char of the current timestep.

int DeepSpeechClient.Models.TokenMetadata.Timestep: Position of the character in units of 20ms.

float DeepSpeechClient.Models.TokenMetadata.StartTime: Position of the character in seconds.

DeepSpeech Interface¶

interface IDeepSpeech¶

Client interface for DeepSpeech

Subclassed by DeepSpeechClient.DeepSpeech

Public Functions

unsafe string DeepSpeechClient.Interfaces.IDeepSpeech.Version(): Return version of this library. The returned version is a semantic version (SemVer 2.0.0).

unsafe int DeepSpeechClient.Interfaces.IDeepSpeech.GetModelSampleRate()

Return the sample rate expected by the model.

Return: Sample rate.

unsafe uint DeepSpeechClient.Interfaces.IDeepSpeech.GetModelBeamWidth()

Get beam width value used by the model. If SetModelBeamWidth was not called before, will return the default value loaded from the model file.

Return: Beam width value used by the model.

unsafe void DeepSpeechClient.Interfaces.IDeepSpeech.SetModelBeamWidth(uint aBeamWidth)

Set beam width value used by the model.

Parameters

aBeamWidth: The beam width used by the decoder. A larger beam width value generates better results at the cost of decoding time.

Exceptions

ArgumentException: Thrown on failure.

unsafe void DeepSpeechClient.Interfaces.IDeepSpeech.EnableExternalScorer(string aScorerPath)

Enable decoding using an external scorer.

Parameters

aScorerPath: The path to the external scorer file.

Exceptions

ArgumentException: Thrown when the native binary failed to enable decoding with an external scorer.
FileNotFoundException: Thrown when cannot find the scorer file.

unsafe void DeepSpeechClient.Interfaces.IDeepSpeech.AddHotWord(string aWord, float aBoost)

Add a hot-word.

Parameters

aWord: Some word
aBoost: Some boost

Exceptions

ArgumentException: Thrown on failure.

unsafe void DeepSpeechClient.Interfaces.IDeepSpeech.EraseHotWord(string aWord)

Erase entry for a hot-word.

Parameters

aWord: Some word

Exceptions

ArgumentException: Thrown on failure.

unsafe void DeepSpeechClient.Interfaces.IDeepSpeech.ClearHotWords()

Clear all hot-words.

Exceptions

ArgumentException: Thrown on failure.

unsafe void DeepSpeechClient.Interfaces.IDeepSpeech.DisableExternalScorer()

Disable decoding using an external scorer.

Exceptions

ArgumentException: Thrown when an external scorer is not enabled.

unsafe void DeepSpeechClient.Interfaces.IDeepSpeech.SetScorerAlphaBeta(float aAlpha, float aBeta)

Set hyperparameters alpha and beta of the external scorer.

Parameters

aAlpha: The alpha hyperparameter of the decoder. Language model weight.
aBeta: The beta hyperparameter of the decoder. Word insertion weight.

Exceptions

ArgumentException: Thrown when an external scorer is not enabled.

unsafe string DeepSpeechClient.Interfaces.IDeepSpeech.SpeechToText(short [] aBuffer, uint aBufferSize)

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

Return

The STT result. Returns NULL on error.

Parameters

aBuffer: A 16-bit, mono raw audio signal at the appropriate sample rate (matching what the model was trained on).
aBufferSize: The number of samples in the audio signal.

unsafe Metadata DeepSpeechClient.Interfaces.IDeepSpeech.SpeechToTextWithMetadata(short [] aBuffer, uint aBufferSize, uint aNumResults)

Use the DeepSpeech model to perform Speech-To-Text, return results including metadata.

Return

The extended metadata. Returns NULL on error.

Parameters

aBuffer: A 16-bit, mono raw audio signal at the appropriate sample rate (matching what the model was trained on).
aBufferSize: The number of samples in the audio signal.
aNumResults: Maximum number of candidate transcripts to return. Returned list might be smaller than this.

unsafe void DeepSpeechClient.Interfaces.IDeepSpeech.FreeStream(DeepSpeechStream 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.

unsafe DeepSpeechStream DeepSpeechClient.Interfaces.IDeepSpeech.CreateStream(): Creates a new streaming inference state.

unsafe void DeepSpeechClient.Interfaces.IDeepSpeech.FeedAudioContent(DeepSpeechStream stream, short [] aBuffer, uint aBufferSize)

Feeds audio samples to an ongoing streaming inference.

Parameters

stream: Instance of the stream to feed the data.
aBuffer: An array of 16-bit, mono raw audio samples at the appropriate sample rate (matching what the model was trained on).

unsafe string DeepSpeechClient.Interfaces.IDeepSpeech.IntermediateDecode(DeepSpeechStream stream)

Computes the intermediate decoding of an ongoing streaming inference.

Return

The STT intermediate result.

Parameters

stream: Instance of the stream to decode.

unsafe Metadata DeepSpeechClient.Interfaces.IDeepSpeech.IntermediateDecodeWithMetadata(DeepSpeechStream stream, uint aNumResults)

Computes the intermediate decoding of an ongoing streaming inference, including metadata.

Return

The extended metadata result.

Parameters

stream: Instance of the stream to decode.
aNumResults: Maximum number of candidate transcripts to return. Returned list might be smaller than this.

unsafe string DeepSpeechClient.Interfaces.IDeepSpeech.FinishStream(DeepSpeechStream stream)

Closes the ongoing streaming inference, returns the STT result over the whole audio signal.

Return

The STT result.

Parameters

stream: Instance of the stream to finish.

unsafe Metadata DeepSpeechClient.Interfaces.IDeepSpeech.FinishStreamWithMetadata(DeepSpeechStream stream, uint aNumResults)

Closes the ongoing streaming inference, returns the STT result over the whole audio signal, including metadata.

Return

The extended metadata result.

Parameters

stream: Instance of the stream to finish.
aNumResults: Maximum number of candidate transcripts to return. Returned list might be smaller than this.