search

sophiefy / StellaVC-ModelList

Model list for the project Stella Voice Changer
GNU General Public License v3.0
5 stars 0 forks source link

Training my own Model #1

Open rx-fly opened 2 years ago

rx-fly commented 2 years ago

What steps would I need to take to train my own voice model? I have audio files/clips I can use, but I am unfamiliar with the process of training an AI Voice model. I've trained a stable diffusion textual inversion model on the webui, but am unsure how different this process would be.

sophiefy commented 2 years ago

What steps would I need to take to train my own voice model? I have audio files/clips I can use, but I am unfamiliar with the process of training an AI Voice model. I've trained a stable diffusion textual inversion model on the webui, but am unsure how different this process would be.

Hi~ Training a voice conversion model is almost the same with the process you train a stable diffusion model. But there are, of course, some tiny differences. I think a good starting point could be StarGANv2-VC, which is totlally self-supervised and you only need to feed it with audio data. The author has also provided detailed training and inference procedure so it won't be difficult to get started:)