Description

First attempt at making model based on anime, will probably replace this with a higher quality model in the future. Trained on 18.5 minutes of audio extracted from the show.