SeedTTSClone<T>

Models & Types · AiDotNet.TextToSpeech.VoiceCloning

Voice-cloning facade over Seed-TTS’s audio/text in-context continuation pipeline.

Inheritance: NeuralNetworkBase<T> → TtsModelBase<T> → SeedTTS<T> → SeedTTSClone<T>

Implements: IVoiceCloner<T>

How It Works

Seed-TTS voice cloning is not a second neural architecture. This class deliberately inherits the same speech tokenizer, causal token LM, token diffusion model, timbre conditioner, and acoustic vocoder. It only adapts the general Seed-TTS prompt API to IVoiceCloner.

Constructors

Constructor Summary
SeedTTSClone(NeuralNetworkArchitecture<T> architecture, SeedTTSCloneOptions options, IGradientBasedOptimizer<T, Tensor<T>, Tensor<T>> optimizer) Creates the facade over the native Seed-TTS pipeline.
SeedTTSClone(NeuralNetworkArchitecture<T> architecture, string modelPath, SeedTTSCloneOptions options) Creates the facade over the native Seed-TTS pipeline.

Properties

Property Summary
double MinReferenceDuration { get; }
int SpeakerEmbeddingDim { get; }

Methods

Method Summary
ModelMetadata<T> GetModelMetadata()
Tensor<T> ExtractSpeakerEmbedding(Tensor<T> referenceAudio)
Tensor<T> SynthesizeWithVoice(string text, Tensor<T> referenceAudio)
Tensor<T> SynthesizeWithVoice(string text, string referenceText, Tensor<T> referenceAudio) Runs the report’s full text-and-audio in-context continuation path.