we up

train to 500 epochs

cause i hear its still slightly dirty

other than that it's pretty good

it's trained off of dirty acapellas lmao

training another 200 epochs isn't gonna fix that

you hotdropped the songs in uvr and called it a day?

lol

no lmfao

the vocal loss after isolation, kareoke and deverb + de echo is crazy

there's only so much I can do when this mfs vocals are buried in effects and crazy bass <:droidsobpixel:900162905402654740>

ikr

i tried to preserve as much as i could and it turned out SHIT

i had to use a criminally low amount of reference amount

and it still sounded dogwater

Ty 🤞🤞

make a v2 with an updated dataset

its crazy

Yeat (RVC) 300 Epoch

its not the best.... but its probably the best on this website so idk

good