SpeechT5Stage
Enums · AiDotNet.TextToSpeech.MultiModal
Which of SpeechT5’s objectives SpeechT5 trains.
Fields
| Field | Summary |
|---|---|
SpeechT5Stage Pretraining |
The joint pre-training (Ao et al. |
SpeechT5Stage TextToSpeech |
Text-to-speech fine-tuning (§3.4): the speech decoder’s L1, the stop token’s BCE and the guided attention loss on recordings and their text. |
int value__ |