I'm trying to make a voice but it won't work.

explain.

LIke what's the error?

it says that my filelist for the transcriptions don't exist, even though it's a folder on the drive

No, it can't be on the drive.

It should be in the filelists folder inside the tacotron2 directory on the colab.

I don't know where that is.

like, i'm not sure where the directory is to move my transcription there

I'm trying to figure out whether tacotron even downloaded or if it didn't

because I cant find that folder anywhere in the colab

what does you're files panel look like

in the drive

not in drive

go in colab and take a screenshot of the files panel

click the triangle next to tacotron2

and there should be a filelists folder

that is where you put your transcription

@Awesome Face (GTX 1660 Ti) tutorial update request: in the vocal remover notebook, the settings at the bottom are optimized for making instrumentals, not vocals there's a beta model optimized for vocals (see replied-to video for a demo), here's how to use it: - set

pretrained_model

to

MGM-v5-Vocal_2Band-32000-BETA1.pth

- set

parameter

to

2band_32000.json

- crank up the aggressiveness, the creator of the notebook uses 0.5 but it might depend on the song - ❗ download the file named "instrumental" which actually has the vocals, they're swapped for technical reasons

oh now i just NEED to try doing one of my datasets again but with that model

its a good thing i kept the original wavs before processing

thanks for the reminder, i should start saving my labels in case vocal remover 6 drops

did it worked like that?

i usually go as notebook says lol

audacity will make me shit my pants in 5 seconds i drag and drop an aup file into the window and it tells me its an audacity project and its telling me to open it from the file menu

i know it is an audacity project. just open the file you do not need to send me into a shitfest just to open an audacity project.

so did i, took me until today to connect the dots

I finished creating the voice model, and it sounds great

Thanks for sharing

trying to do a danganronpa voice, whats the recommended amount of samples?

hey folks doing a tacotron2 voice based on the YouTube video guide and getting an error running the training on Colab. it looks like utils.py is calling filelists/clipper_train_filelist.txt, but the file doesn't exist. it's pretty much blocking it from running. i don't know anything about this file. anyone got a clue? or maybe a link to a better Colab tacotron2 example? any advice appreciated!