模型:

speechbrain/tts-hifigan-libritts-16kHz