Description

< Created by me (AI_Characters). Please @ my Twitter or Instagram accounts when using my model. I would love to know when it is used! If you want to support what I am doing, donations to my Ko-Fi ( would be greatly appreciated! Those will go towards funding the renting of GPUs for further experimentation and model making. Dataset is from a 4h YouTube video of just his ingame voicelines, cut down to around 15 minutes of audio. With editing and silence-truncating it is down to 12 minutes of pure audio of him. Does very well for speech, and worse for songs. He doesn't sing and his voice is very deep. I recommend keeping search feature ratio at or close to 0 for singing. Inference settings used for these samples, except for 0.5 feature search ratio for the speech sample: Mangio-Crepe, 64 hop length, 3 median filter, 0.2 volume envelope, 0.1 protect Any helpful and honest critique of the model is greatly appreciated!