A demo of tada-3b-ml — a text-to-speech model that clones voice, emotion, and timing from a short audio prompt.
How to use: Choose a voice prompt (or upload your own), enter text, and click Generate. The model will encode the prompt and generate speech in one step.