StyleTtsTrainingStage
Enums · AiDotNet.TextToSpeech.StyleEmotion
The modules a StyleTTS training step updates (Li et al. 2022, §II-B).
Fields
| Field | Summary |
|---|---|
StyleTtsTrainingStage Prosody |
Stage 2: the duration and prosody predictors, with every other module fixed. |
StyleTtsTrainingStage Reconstruction |
Stage 1: mel reconstruction — the text encoder, style encoder, decoder and, once the transferable monotonic aligner’s objectives start, the text aligner and pitch extractor. |
int value__ |