It requires some knowledge of the language you're building the bank in tbh

Because you have to know how the words should be pronounced correctly to be able to teach the program to then know how to pronounce them

What I might also suggest is to pay attention to the default Bank. I think it's Teto for Windows and it's Defoko for Mac, but take a look at their bank settings and see how things are set up there, that'll also give you a good idea of where to go And as for using melodyne I have to recommend against it because of the issue with altering the quality of the audio which auto-tune does, it may very well end up hurting the quality of the bank even further

( I don't know who you have downloaded or if collecting UTAUs is a thing you're into but I HIGHLY recommend Ritsu bc he has a gorgeous vcv bank and if I remember correctly, even comes in a less complicated light list version you can pull your reclist from )

They could of dug through the tankmen series but I'm guessing they didn't.

Idk if it belongs here but I made AI Wrld british

Checking Training Files wav/1.npy| [WARNING] does not exist. --------------------------------------------------------------------------- FileNotFoundError Traceback (most recent call last) in () ----> 1 check_dataset(hparams) 2 frames /usr/local/lib/python3.7/dist-packages/numpy/lib/npyio.py in load(file, mmap_mode, allow_pickle, fix_imports, encoding) 414 own_fid = False 415 else: --> 416 fid = stack.enter_context(open(os_fspath(file), "rb")) 417 own_fid = True 418 FileNotFoundError: [Errno 2] No such file or directory: 'wav/1.npy'

wav/1.wav doesn't exist

check your dataset man

use #841732367962210344 channel

Alright, i didnt copy the whole thing though there is text there

pffft..

how to train songs with my own voice

May be you put an extra "/" in front of "wav".

Finally got the colab working, how long does it take for a model to train?

Error 503

It's not just Github. Larger sites like Reddit have that problem too right now.

try again

Yeah, same error

is this good at least?

not really

could be worse

here it works 🤷

huh, this is the one immediately after the last image I sent

that's as good one can expect

and now it's back to what it was before

there is some randomness in process

you should stop training soon