A lightweight text-to-speech model from the Qwen3 family optimized for custom voice cloning with 12Hz audio output.
No resource links recorded.