Description
Huggingface: Weights.gg: Version firev2: RVC V2 - Ov2 Super Pretrain Pitch Extraction Algorithm : Rmvpe Epochs : 400 Dataset Length : ~ 00:06:59 minutes of dataset extracted from all the speaking voicelines of the character. Model created by @Voxel. If you can, please credit me when you use the model.
Comments

did u use the english ver or the japanese ver?

yes

i heard it

its not bad

I remember months ago there was a hu tao model with 58 minutes dataset ripped directly from the game

yeah, it was before ai hub was shut down, idk where that person is now haha

idk it was ripped directly from the game he says

after that he made this


fr

this is the cleanest and smoothest model that ever existed

yep, no robotic sibilance, breathes or what ever voice cracks

that was the time it's stuck in my head that 1 hour is really the sweet spot for a natural sounding model

ShyElijah, I think.

He was the Genshin model guy.

yeah thats me. Probably over an hour now if i bothered re-ripping it. Reason why is cause i also had story quest voicelines
hola cómo estás
Add a comment
Samples
Pitch