OptionalallowAllow preview and experimental models.
You confirm the speaker agreed to have their voice cloned. Required.
OptionaldeviceForce a device. Default 'auto'.
OptionaldtypeForce a dtype, such as 'q4'.
OptionalmodelA cloning model. Default 'voice:clone'.
OptionalonReceives load events while the model downloads and compiles.
3 to 10 seconds of clear speech from one speaker.
OptionalsampleSample rate of reference when it is raw samples.
OptionalsaveKeep the voice for later as voice: 'saved:<name>'.
OptionalsignalAbort download or inference.
Options shared by every verb.