Languageisteğe bağlı
Text-normalization language code. Supported: en_us, en_gb, fr_fr, de, es, it, pt_br, ja, cmn, ko. Unsupported codes skip normalization.
Seedisteğe bağlı
Seed for reproducibility. Random when omitted.
Accurate Modeisteğe bağlı
True = closer voice match; False = more expressive.
Clean Speaker Backgroundisteğe bağlı
Mark the reference audio as having a clean background.
Temperatureisteğe bağlı
Sampling temperature.
Top Pisteğe bağlı
Nucleus sampling probability (0 disables).
Min Pisteğe bağlı
Minimum-probability sampling threshold.
Top Kisteğe bağlı
Top-k sampling cutoff (0 disables).
Max Tokensisteğe bağlı
Maximum number of audio frames to generate. Defaults to the model context limit.
Enable Safety Checkeristeğe bağlı
Authorized callers may set false to request relaxed screening.
