PlayVoice / vits_chinese

Best practice TTS based on BERT and VITS with some Natural Speech Features Of Microsoft; Support ONNX streaming out!
https://huggingface.co/spaces/maxmax20160403/vits_chinese
MIT License
1.16k stars 168 forks source link

关于ZeroShot #60

Closed suzhenghang closed 1 year ago

suzhenghang commented 1 year ago

大佬有打算后续做下ZeroShot的工作嘛?或者有了解过ZeroShot目前性能怎样啊?

MaxMax2016 commented 1 year ago

这个ZeroShot都是靠语料和算力堆起来的,这个~~做不了啊

TinaChen95 commented 1 year ago

是不是可以用 external speaker embedding + aishell3 试一试? 看到另一个讨论里面,多说话人效果变差。不太确定这个变差是差了多少。 想请教下算力上预估会需要多少,以及您是怎么预估的呢?

scriptboy1990 commented 11 months ago

这个ZeroShot都是靠语料和算力堆起来的,这个~~做不了啊

可以进一步解释下吗,为啥是靠算力推起来的