SparkTTSTrainingComponent
Enums · AiDotNet.TextToSpeech.CodecBased
The independently trainable stages of Spark-TTS.
Fields
| Field | Summary |
|---|---|
SparkTTSTrainingComponent BiCodec |
BiCodec reconstruction, semantic VQ and speaker-token objectives. |
SparkTTSTrainingComponent LanguageModel |
Qwen2.5 causal next-token prediction over text, global speaker tokens and semantic tokens. |
int value__ |