How do I load a dataset into RVC Mainline Colab?

[hollow jewel]

I was trying to load a dataset to start preprocess for training a voice model and I got an error.

[zinc goblet]

try using the google drive path instead

[hollow jewel]

I already did using "/content/training/datasets" and it still gives me this error.

[fair imp]

make sure the dataset is zipped properly, i.e. containing a folder with the same name as the model and containing wav files

[hollow jewel]

That's what I did. I had the dataset zipped with the folder containing the same name as the model and containing .wav files. I even typed the name of the model I'm training case sensitive. Again, it's still giving me the same error as before.

[hollow jewel]

Here's the code (note: I have to provide it in .txt file format due to Discord's character limit):

[zinc goblet]

zipped datasets dont work on mainline

its needs to be raw audios

[hollow jewel]

Zipped datasets don't work on Mainline? Then what, do I have to upload the files exactly as exported? No folders necessary?

[zinc goblet]

you need to have it as extracted in your google drive

[hollow jewel]

Okay. I'll probably try that then.

[solid mortar]

So we don't have to place it in the Imjoy Elfinder GUI and can place the audios here?

[zinc goblet]

yes

[fair imp]

idk though it's true for local one

[hollow jewel]

Okay. I just uploaded the folder from the dataset's .zip file onto Google Drive in the dataset folder, copied the path of the audio files and it works. No error.

[solid mortar]

oh good. That would be hell since you know the whole audio label method

[hollow jewel]

Definitely. Thanks.

[solid mortar]

no thank you for asking. Just jotting this down

[main anvil]

yo

it says "place audio here" but nothing happens when i upload audio

on ngrok

how can i fix that