WebKit

Discover the powerful RVC v2 Model by Wendigoon, specifically trained for creating AI music! This version has undergone 325/950 epochs, offering unique and customizable sounds generated from a dataset of 43 hours worth of training data from Harvest. With the ability to start overtraining around the 18,400 step mark, this model gives you the flexibility to choose the sound that suits your needs best. Dive into the world of artificial intelligence music creation and explore endless possibilities with Wendigoon's RVC v2 Model.

Dataset was 43:03: Trained on Harvest. It started overtraining at about 18,400 steps, but I couldn't decide which checkpoint sounded better, so I included them both. You can decide for yourself.

Wendigoon 325

Wendigoon (RVC v2) 325 / 950 Epoch

analogspiderweb

idk if it'll even let me with what little audio there is

Tux (Tux Paint)

Dataset was 1:05:52. Echo removed & audio balanced using CrumplePop.

Wheatley HD (Portal 2)

Wheatley HD (Portal 2) (RMVPE) (RVC v2) 450 Epoch

Dataset was 59:46. Echo was removed with CrumplePop:

Wendigoon 325

Description

Samples

Wendigoon