So Vits SVC 4.0 offline training (a very derped method) - PPP 2023

doesnt lowering batch side degrade the model

Will there be a new proper tutorial?

thank you for this. will you do a new tutorial for offline vocal synthesis for 4.0 with clustering enabled?

Please tell me why Diff-SVC training doesn't work? Colab

nice tutorial! I can do walkthrough this video. I got 2 question.
1. How should i know when model finish traning? can stop around G_20000?
2. When i finished train and get .pth model. how i use it to change music singer voice?

Is it better to train the voices with different emotions, or just one?

4.0 in colab tut next

Thanks for thís tutorial.
Can you tell me how can i train another model after i finished one,
or continue to train a previous model?

Do you think you could help me make my own model

What about training from command/anaconda prompt instead of colab? Its 4 python commands after putting the dataset in dataset_raw so i think its easier than using the colab locally due to it having less steps.

hello amoawesomeart ive downloaded your tree hugger model off hugging face but when i launch so vits 4.0 with it, it says there is no clustering model and it wont launch now ?

for some reason ive downloaded a bunch of models for my sovits svc 4.0 and it wont initialize some models it thinks g_0 and D_0 pth is missing and it says skipping [Models] in lines how do i fix this ?

Skipping Fluttershy Speaking [Mix], no G_*.pth
Skipping Maud Pie [Speaking], no G_*.pth
Skipping Princess Celestia [Speaking], no G_*.pth
Skipping Princess Luna [Singing], no G_*.pth
Skipping Princess Luna [Speaking], no G_*.pth
Skipping Scootaloo [Singing], no G_*.pth

How to fix this issue ?

If you happen to get this error when installing the dependencies: "ERROR: Failed building wheel for fairseq ERROR: Could not build wheels for fairseq, which is required to install pyproject.toml-based projects"...,
-Make sure you install C++ Build Tools.

does it really need to be a nvidia graphics card??

where is the link to the offline training colab ?

hey man can you make a fresh tutorial for how to use 4.0 because ive been sitting here on my computer researching for 3 days and its still not working for me i just dont know what im doing wrong

When getting error that looks like this: 'ValueError: n_samples=598 should be >= n_clusters=4000. 0 0.583984375 MB , shape: (598, 256) float32'
Check if your dataset is completely made out of wav files, Ive just come across this error and turned out my dataset was a mix of mp3 and wav files.

where download folderDumpText/offlinesovit/sovits_EN_training_colab_4_0_offline_(231703).ipynb ???

If I started in google colab and reached 20 thousand steps and now I want to go offline, how can I do that?

[WinError 3] The system cannot find the path specified: '/content/so-vits-svc

i would like to mention that RVC is out Now Are you able to provide a installation tutorial on RVC

There is 3 Trained models for it already

what i need to do with that massage in google colab 'git' is not recognized as an internal or external command,
operable program or batch file.

Can I do it using Intel GPU?

Is there any way to make it work without using a GPU?

Hi, any way you can think to do the voice changing in real time?

Warning, this is NOT a proper tutorial, just merely my way of training offline models for So-Vits-SVC 4.0, so this method may work on your pc or may not, due to mysterious ways of google colab and python constant breaking compatibility of modules so I would be prepared to not be surprised if this code do not work on your machine.