F5-TTS, the new state-of-the-art in open source voice cloning
This model is cold. You'll get a fast response if the model is warm and already running, and a slower response if the model is cold and starting up.
