OpenVoiceTrainingOptions
Options & Configuration · AiDotNet.TextToSpeech.VoiceCloning
Paper section 2.3 training graph. Loss/optimizer settings not published by OpenVoice use configurable VITS/HiFi-GAN baseline choices, not claimed recovered training settings.
Constructors
| Constructor | Summary |
|---|---|
OpenVoiceTrainingOptions() |
Creates an instance with the default settings. |
OpenVoiceTrainingOptions(OpenVoiceTrainingOptions other) |
Copies the settings from another options instance. |
Properties
| Property | Summary |
|---|---|
bool IncludeTrainingComponents { get; set; } |
Builds the auxiliary prior/discriminators before first training; training enables and retains this flag automatically. |
double Beta1 { get; set; } |
Gets or sets the first-moment decay for the training optimizer. |
double Beta2 { get; set; } |
Gets or sets the second-moment decay for the training optimizer. |
double Epsilon { get; set; } |
Gets or sets the numerical stability constant of the optimizer. |
double FeatureMatchingWeight { get; set; } |
Gets or sets the weight of the feature matching objective. |
double IpaDropout { get; set; } |
Gets or sets the dropout probability for the IPA encoder. |
double KlWeight { get; set; } |
Gets or sets the weight of the kl objective. |
double LearningRate { get; set; } |
Gets or sets the optimizer learning rate. |
double LearningRateDecay { get; set; } |
Gets or sets the learning rate decay. |
double MelLossWeight { get; set; } |
Gets or sets the weight of the mel loss. |
double MelMaxFrequency { get; set; } |
Gets or sets the mel max frequency. |
double WeightDecay { get; set; } |
Gets or sets the optimizer weight decay. |
int DiscriminatorWidthDivisor { get; set; } |
One retains all discriminator channels. |
int IpaFilterChannels { get; set; } |
Gets or sets the IPA encoder filter channels. |
int IpaHeads { get; set; } |
Gets or sets the number of attention heads in the IPA encoder. |
int IpaKernel { get; set; } |
Gets or sets the IPA encoder kernel. |
int IpaLayers { get; set; } |
Gets or sets the number of layers in the IPA encoder. |
int IpaVocabularySize { get; set; } |
Size of the caller’s universal IPA phoneme-token inventory. |
int MelChannels { get; set; } |
Gets or sets the mel channels. |
int SamplingSeed { get; set; } |
Gets or sets the random seed used during sampling. |
int SegmentFrames { get; set; } |
Gets or sets the segment frames. |
int UpdatesPerEpoch { get; set; } |
Gets or sets the updates per epoch. |
long TrainingSteps { get; set; } |
Completed utterance steps, retained with the options for draw/schedule continuation. |