fixed speaker synthesis without prompt

[HYJ-creator]

Hello, can ZipVoice support fixed speaker synthesis now? For example, like FastSpeech1/2, where you don't need a prompt but can specify the timbre via a speaker ID

[zhu-han]

Our pre-trained model does not support it.