KotobaWhisper<T>
Models & Types · AiDotNet.SpeechRecognition.WhisperFamily
Kotoba-Whisper: Distil-Whisper’s recipe applied to Japanese, a large-v3 encoder with a 2-layer decoder.
Inheritance: NeuralNetworkBase<T> → AudioNeuralNetworkBase<T> → KotobaWhisper<T>
Implements: ISpeechRecognizer<T>
For Beginners
Kotoba-Whisper listens to 16 kHz audio in 30-second windows turned into a log-mel spectrogram, encodes each window with a Transformer, and a second Transformer writes the transcript token by token while reading the encoded audio.
How It Works
Kotoba Technologies distilled large-v3 on the ReazonSpeech corpus with Distil-Whisper’s method (Gandhi et al. 2023): the teacher’s 32-layer, 1280-wide encoder with a 2-layer decoder, Whisper’s tokenizer, and Japanese as the default language. Every forward reads the 30-second window openai-whisper builds (pad_or_trim); Predict returns the
decoder’s logits over the start sequence (start of transcript, language, transcribe, no timestamps), whose last row is
the first token’s distribution, and Train minimizes the teacher-forced cross-entropy of a transcript given as
token ids or as one distribution per token.
Example
var architecture = new NeuralNetworkArchitecture<float>(
inputType: InputType.OneDimensional,
taskType: NeuralNetworkTaskType.SpeechRecognition,
inputSize: 480000, outputSize: 51866);
var model = new KotobaWhisper<float>(architecture);
var text = model.Transcribe(waveform).Text;
Constructors
| Constructor | Summary |
|---|---|
KotobaWhisper(NeuralNetworkArchitecture<T> architecture, KotobaWhisperOptions options, IGradientBasedOptimizer<T, Tensor<T>, Tensor<T>> optimizer) |
|
KotobaWhisper(NeuralNetworkArchitecture<T> architecture, string modelPath, KotobaWhisperOptions 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 16 kHz audio with openai-whisper’s transcribe: 30-second windows, temperature fallback, and word timing when timestamps are requested. |
string DetectLanguage(Tensor<T> audio) |