1 paper · 1 filter
John Zheng, Farhad Maleki
In zero-shot text-to-speech, achieving a balance between fidelity to the target speaker and adherence to text content remains a challenge. While classifier-free guidance (CFG) stra…