RVC-Project/Retrieval-based-Voice-Conversion-WebUI

Fork 0

mirror of https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI.git synced 2026-11-05 00:27:30 +00:00

[GH-ISSUE #1211] Perhaps it's time to find an alternative to google colab? #783

New issue

Open

opened 2026-10-20 10:03:00 +00:00 by bsakaguchi · 17 comments

bsakaguchi commented

2026-10-20 10:03:00 +00:00

Owner

Originally created by @ShiroHusky on GitHub (Sep 9, 2023).
Original GitHub issue: https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/issues/1211

They sure are trying to force us to get colab Pro by disconnecting every 6 minutes or so, but as the good Brazilian I am, I ain't paying a kidney for some computing unit that ends in two days

So, perhaps try finding an alternative to Colab that could run RVC/Easygui?
Or you know, solve the "disconnect issue" it on the good ole "Jeitinho Brasileiro" if you know what I mean

Originally created by @ShiroHusky on GitHub (Sep 9, 2023). Original GitHub issue: https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/issues/1211 They sure are trying to force us to get colab Pro by disconnecting every 6 minutes or so, but as the good Brazilian I am, I ain't paying a kidney for some computing unit that ends in two days So, perhaps try finding an alternative to Colab that could run RVC/Easygui? Or you know, solve the "disconnect issue" it on the good ole "[Jeitinho Brasileiro](https://en.wikipedia.org/wiki/Jeitinho)" if you know what I mean

bsakaguchi added the

todo list

help wanted

labels

2026-10-20 10:03:00 +00:00

bsakaguchi commented

2026-10-20 10:03:01 +00:00

Author

Owner

@NatoBoram commented on GitHub (Sep 9, 2023):

You could always run it locally, it has CPU training and conversion support

@NatoBoram commented on GitHub (Sep 9, 2023): You could always run it locally, it has CPU training and conversion support

bsakaguchi commented

2026-10-20 10:03:01 +00:00

Author

Owner

@ShiroHusky commented on GitHub (Sep 9, 2023):

You could always run it locally, it has CPU training and conversion support

How do I run it locally, is there any tutorials for that? I'm Not tech savvy and all x3

@ShiroHusky commented on GitHub (Sep 9, 2023): > You could always run it locally, it has CPU training and conversion support How do I run it locally, is there any tutorials for that? I'm Not tech savvy and all x3

bsakaguchi commented

2026-10-20 10:03:01 +00:00

Author

Owner

@loboere commented on GitHub (Sep 11, 2023):

@NatoBoram not all of us have nasa computers

@loboere commented on GitHub (Sep 11, 2023): @NatoBoram not all of us have nasa computers

bsakaguchi commented

2026-10-20 10:03:01 +00:00

Author

Owner

@ShiroHusky commented on GitHub (Sep 11, 2023):

@NatoBoram not all of us have nasa computers

I mean, my gaming PC is took about two hours just to get out of the first epoch, despite having a Ryzen5 with built in graphics, if a gaming device takes this long, then imagine PC's with lower specs

@ShiroHusky commented on GitHub (Sep 11, 2023): > @NatoBoram not all of us have nasa computers I mean, my gaming PC is took about two hours just to get out of the first epoch, despite having a Ryzen5 with built in graphics, if a gaming device takes this long, then imagine PC's with lower specs

bsakaguchi commented

2026-10-20 10:03:01 +00:00

Author

Owner

@NatoBoram commented on GitHub (Sep 11, 2023):

It depends on the size of your dataset. If you have 1h of data, it's not feasible on a CPU, but if you have 10m, then one epoch might take 5 minutes and then you can train overnight.

@NatoBoram commented on GitHub (Sep 11, 2023): It depends on the size of your dataset. If you have 1h of data, it's not feasible on a CPU, but if you have 10m, then one epoch might take 5 minutes and then you can train overnight.

bsakaguchi commented

2026-10-20 10:03:01 +00:00

Author

Owner

@instant-high commented on GitHub (Sep 11, 2023):

I'm playing around with an old GTX 1060/3gb
2 min. Dataset is 2:30 min. per epoch.
Testing how far I can go with the dataset length

@instant-high commented on GitHub (Sep 11, 2023): I'm playing around with an old GTX 1060/3gb 2 min. Dataset is 2:30 min. per epoch. Testing how far I can go with the dataset length

bsakaguchi commented

2026-10-20 10:03:01 +00:00

Author

Owner

@RVC-Boss commented on GitHub (Sep 13, 2023):

I think finding a alternative to colab is a necessary thing. If you have any good suggestions, you can also let me know.

@RVC-Boss commented on GitHub (Sep 13, 2023): I think finding a alternative to colab is a necessary thing. If you have any good suggestions, you can also let me know.

bsakaguchi commented

2026-10-20 10:03:01 +00:00

Author

Owner

@blaisewf commented on GitHub (Sep 13, 2023):

Kaggle, Paperspace or improve code and promote local should be the best options!

@blaisewf commented on GitHub (Sep 13, 2023): Kaggle, Paperspace or improve code and promote local should be the best options!

bsakaguchi commented

2026-10-20 10:03:02 +00:00

Author

Owner

@Tps-F commented on GitHub (Sep 13, 2023):

I am currently working on this.

@Tps-F commented on GitHub (Sep 13, 2023): I am currently working on this.

bsakaguchi commented

2026-10-20 10:03:02 +00:00

Author

Owner

@loboere commented on GitHub (Sep 23, 2023):

any news?

@loboere commented on GitHub (Sep 23, 2023): any news?

bsakaguchi commented

2026-10-20 10:03:02 +00:00

Author

Owner

@MRINFINITE0 commented on GitHub (Oct 6, 2023):

Greetings! @Tps-F @RVC-Boss I have an suggestion to implement notebook on azure vm or azure notebook , I tried to implement there but stuck on FFMPEG error while processing data in train model section can we work around azure part or is there error from my side while copying code from colab to azure notebook .... please reply
Thanks in advance

@MRINFINITE0 commented on GitHub (Oct 6, 2023): Greetings! @Tps-F @RVC-Boss I have an suggestion to implement notebook on azure vm or azure notebook , I tried to implement there but stuck on FFMPEG error while processing data in train model section can we work around azure part or is there error from my side while copying code from colab to azure notebook .... please reply Thanks in advance

bsakaguchi commented

2026-10-20 10:03:02 +00:00

Author

Owner

@Tps-F commented on GitHub (Oct 6, 2023):

I see, I didn't know that service existed.
I'll give it a try.

@Tps-F commented on GitHub (Oct 6, 2023): I see, I didn't know that service existed. I'll give it a try.

bsakaguchi commented

2026-10-20 10:03:02 +00:00

Author

Owner

@MRINFINITE0 commented on GitHub (Oct 6, 2023):

@Tps-F it doesnt exist we have implemented that

@MRINFINITE0 commented on GitHub (Oct 6, 2023): @Tps-F it doesnt exist we have implemented that

bsakaguchi commented

2026-10-20 10:03:02 +00:00

Author

Owner

@pursuitcurves commented on GitHub (Oct 11, 2023):

try running it on a hosted runtime, i do not get disconnected and the ui link stays active for atleast 2 days.

https://research.google.com/colaboratory/marketplace.html

They have some issues with mounting drive (https://github.com/googlecolab/colabtools/issues/2533), so create a fake /content/drive folder and keep everything you do there.

If during training, you might face "ran out of shared memory error", follow the first answer here (https://stackoverflow.com/questions/58804022/how-to-resize-dev-shm) and add 20G of shared memory.

Cheers!

@pursuitcurves commented on GitHub (Oct 11, 2023): try running it on a hosted runtime, i do not get disconnected and the ui link stays active for atleast 2 days. https://research.google.com/colaboratory/marketplace.html They have some issues with mounting drive (https://github.com/googlecolab/colabtools/issues/2533), so create a fake /content/drive folder and keep everything you do there. If during training, you might face "ran out of shared memory error", follow the first answer here (https://stackoverflow.com/questions/58804022/how-to-resize-dev-shm) and add 20G of shared memory. Cheers!

bsakaguchi commented

2026-10-20 10:03:02 +00:00

Author

Owner

@Hedwig7s commented on GitHub (Oct 28, 2023):

Kaggle, Paperspace or improve code and promote local should be the best options!

Paperspace has capacity issues on free tier and Kaggle seems to have blacklisted WebUI similar to Colab

@Hedwig7s commented on GitHub (Oct 28, 2023): > Kaggle, Paperspace or improve code and promote local should be the best options! Paperspace has capacity issues on free tier and Kaggle seems to have blacklisted WebUI similar to Colab

bsakaguchi commented

2026-10-20 10:03:02 +00:00

Author

Owner

@Hedwig7s commented on GitHub (Oct 28, 2023):

For Kaggle or Colab to work without paying WebUI needs to be completely removed from it (trying to hide it will result in an account termination don't try)

@Hedwig7s commented on GitHub (Oct 28, 2023): For Kaggle or Colab to work without paying WebUI needs to be completely removed from it (trying to hide it will result in an account termination don't try)

bsakaguchi commented

2026-10-20 10:03:02 +00:00

Author

Owner

@Hedwig7s commented on GitHub (Dec 3, 2023):

Any updates on this?
Also for my above comment: https://github.com/googlecolab/colabtools/issues/3996#issuecomment-1719793880

@Hedwig7s commented on GitHub (Dec 3, 2023): Any updates on this? Also for my above comment: https://github.com/googlecolab/colabtools/issues/3996#issuecomment-1719793880