AdaSpeech2Options
Options & Configuration · AiDotNet.TextToSpeech.Classic
Options for AdaSpeech 2 (adaptive TTS with untranscribed speech data).
Inheritance: ModelOptions → TtsModelOptions → AcousticModelOptions → FastSpeech2Options → AdaSpeechOptions → AdaSpeech2Options
For Beginners
These options configure the AdaSpeech2 model. Default values follow the original paper settings.
How It Works
AdaSpeech 2’s TTS pipeline is AdaSpeech’s (Yan et al. 2021, §2.1, §3.1: hidden 256, 2 heads, FFN filter 1024
and kernel 9, 80-bin mel, 16 kHz audio with a 12.5 ms hop), so these options extend AdaSpeechOptions
with the mel-spectrogram encoder.
Constructors
| Constructor | Summary |
|---|---|
AdaSpeech2Options() |
Creates the paper’s configuration. |
AdaSpeech2Options(AdaSpeech2Options other) |
Initializes a new instance by copying from another instance. |
Properties
| Property | Summary |
|---|---|
int NumMelEncoderLayers { get; set; } |
Gets or sets the feed-forward Transformer blocks of the mel-spectrogram encoder (4, symmetric with the phoneme encoder, §2.2). |