OmniVoice TTS Model: AMD GPU Version with ROCm Acceleration Features Zero-shot text-to-speech for over 600 languages AMD GPU acceleration (ROCm 7.2+) Voice cloning and voice design RTF as low as 0.06 (16x real-time) System Requirements Python 3.12.x (⚠️ ROCm 7.2.1 supports Python 3.12 only)