←back to thread

VibeVoice: A Frontier Open-Source Text-to-Speech Model

(microsoft.github.io)
448 points lastdong | 4 comments | 03 Sep 25 10:44 UTC | HN request time: 0.855s | source
1. agos ◴[03 Sep 25 14:46 UTC] No.45116442[source]▶
seemingly supports only English, Indian and Chinese
replies(1): >>45117828 #
2. plingamp ◴[03 Sep 25 16:38 UTC] No.45117828[source]▶
Indian and Chinese are not languages
replies(2): >>45119023 #>>45125403 #
3. ascorbic ◴[03 Sep 25 18:33 UTC] No.45119023[source]▶
Voices, not languages. The "English" one is American though.
4. agos ◴[04 Sep 25 09:44 UTC] No.45125403[source]▶
I'm very aware of this. The project does not specify more than an in- and zh- prefix.