Description
Model on only 3m30s of data, ~~might~~ probably will update later with a more trained model with more data but as of yet the only way to rip voices from the game is via line out sample audio is weird al yankovic - weasel stomping day
Comments

Bro stop posting 20e models I'm going to go to your house and force you to train for longer

You have a 3060 12gb you can literally cache the entire dataset to VRAM and train at 8-16 batch size and pump out 200e in like 45 minutes
Add a comment
Samples
Pitch
More models by
ezogaming
Japanese Crested Ibis / Toki (Kemono Friends) (Screaming), 32k
1 year ago0
6
Serval (Kemono Friends) (Yuka Ozaki)
1 year ago2
10
Southern Rockhopper Penguin (Iwabee) (Iwabi) (Rocker) (Kemono Friends)
1 year ago4
Fennec (Kemono Friends)
1 year ago0
3
Common Raccoon (Arai-san) (Kemono Friends)
1 year ago0
7
More to explore
Loading more