High-fidelity Text-To-Speech
F5-TTS & E2-TTS: Zero-Shot Voice Cloning (Unofficial Demo)
watermark-free Modelscope-based video generation