rsxdalv / tts-generation-webui

TTS Generation Web UI (Bark, MusicGen + AudioGen, Tortoise, RVC, Vocos, Demucs, SeamlessM4T, MAGNet, StyleTTS2, MMS)
https://rsxdalv.github.io/tts-generation-webui/
MIT License
1.63k stars 175 forks source link

Train on expressions #94

Closed adhikjoshi closed 1 year ago

adhikjoshi commented 1 year ago

Hi, voice cloning is working good with bark, is there way of training more expressions in Bark?

rsxdalv commented 1 year ago

Hi, perhaps this could help: https://github.com/anyvoiceai/Barkify also, there's some prompt engineering possible (although Bark often doesn't "comply"), so you can get an angry laugh or a silly laugh etc.

MiYakuGaming commented 1 year ago

I've been doing great with Tortoise TTS using [I am very happy and sad,] Your emotional text There's variants that have worked for me such as [I am very happy and yelling with excitement,] Your emotional text

rsxdalv commented 1 year ago

I will close the issue for now. Please reopen it if there are more questions.