Introducing the best John Frusciante RVC V2 model to date.
Trained on rmvpe RVC V2, 32khz, 12 batch size, 540 epochs.
Trained on 17 minutes of high quality extracted vocal tracks from some of Frusciante's best solo works. Audio was hand currated and processed in audacity. This model sounds amazing when only a slight reverb and EQ is added, and tops anything currently publicly listed. All the other models I have found have been extremely robotic, and this is not.
The zip download only includes the model. But in the repo there is a demo wav file (unaltered), and an EQ preset for audacity. I suggest using the EQ preset then adding Vocal I reverb. If you want, you can also do a bass and treble boost BEFORE the EQ preset. Credit to VoiceOverMaster for the EQ Preset.
You can find a demo cover of The Wind on my youtube: https://youtu.be/BG7aQqGA80w
The demo in the repo / zip file is unaltered output from RVC. If you would like the dataset, send me a DM on discord- I dont think I would be able to upload it for copyright.