StreamingConformer<T>

Models & Types · AiDotNet.SpeechRecognition.Streaming

A streaming Conformer transducer: the Conformer of Gulati et al. 2020 with attention limited to fixed chunks.

Inheritance: NeuralNetworkBase<T> → AudioNeuralNetworkBase<T> → StreamingConformer<T>

Implements: ISpeechRecognizer<T>

For Beginners

Streaming Conformer turns 16 kHz audio into an 80-band log-mel spectrogram, shortens it 4x with two convolutions, and encodes it with Conformer blocks, each mixing self-attention (long-range context) with a convolution (local patterns). A small LSTM remembers the words written so far and a joint network combines it with each audio frame to pick the next token or “nothing yet”.

How It Works

The encoder is Conformer (L) of Table 1 with an RNN-T decoder; attention runs within chunks of encoder frames so each output depends on bounded context, the chunk-wise streaming used for Conformer transducers. SpecAugment masks the features in training (section 3.1). The RNN-T decoder (Graves 2012) trains with the transducer loss and decodes greedily, at most ten labels per frame. The paper’s blocks use relative positional encoding; these use rotary position encoding, the relative scheme ConformerBlockLayer provides.

Example

var architecture = new NeuralNetworkArchitecture<float>(
    inputType: InputType.OneDimensional,
    taskType: NeuralNetworkTaskType.SpeechRecognition,
    inputSize: 16000, outputSize: 34);
var model = new StreamingConformer<float>(architecture);
var text = model.Transcribe(waveform).Text;

Constructors

Constructor Summary
StreamingConformer(NeuralNetworkArchitecture<T> architecture, StreamingConformerOptions options, IGradientBasedOptimizer<T, Tensor<T>, Tensor<T>> optimizer)
StreamingConformer(NeuralNetworkArchitecture<T> architecture, string modelPath, StreamingConformerOptions options)

Properties

Property Summary
IReadOnlyList<string> SupportedLanguages { get; }
bool SupportsStreaming { get; }
bool SupportsWordTimestamps { get; }

Methods

Method Summary
IReadOnlyDictionary<string, T> DetectLanguageProbabilities(Tensor<T> audio)
IStreamingTranscriptionSession<T> StartStreamingSession(string language)
ModelMetadata<T> GetModelMetadata()
Task<TranscriptionResult<T>> TranscribeAsync(Tensor<T> audio, string language, bool? includeTimestamps, CancellationToken cancellationToken)
Tensor<T> ForwardForTraining(Tensor<T> input)
TranscriptionResult<T> Transcribe(Tensor<T> audio, string language, bool? includeTimestamps) Transcribes audio: the encoder and the transducer with greedy decoding.
string DetectLanguage(Tensor<T> audio)