I've been trying to make a model of Kaworu's voice in Latin Spanish but it always ends up sounding weird and I have no idea if it's the dataset or he just lacks training.
So I am attaching the dataset that I took from the movie "Evangelion: 3.33 You Can (Not) Redo" so that you only have to train it.
https://drive.google.com/file/d/1dQHKU6g7mGcbXEXqZi20apAwXrN0bytk/view?usp=sharing
Google Docs