Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Would it be possible to use Google colab to train your own dataset? #160

Open
Zylvian opened this issue Oct 9, 2020 · 6 comments
Open

Comments

@Zylvian
Copy link

Zylvian commented Oct 9, 2020

I mostly wanna use Jukebox for trying to finish my own songs, and I guess I would have to train my own dataset on my own songs. I have a pretty ass laptop atm, would it be possible to train my own dataset with my own songs using Colab?

@michaelklachko
Copy link

no

@Zylvian
Copy link
Author

Zylvian commented Oct 10, 2020

@michaelklachko Well dangit. Do you have any estimates on the performance on a single 1070, as well as how much I would actually get out of it, seeing as my band only has 6 songs?

Furthermore, any idea how it'll fare in Norwegian?

@Apokar
Copy link

Apokar commented Oct 14, 2020

@Zylvian I'm trying to generate music in Chinese, But I stuck in training vqvae,lol. let me know if you have any progress, Thank you

@michaelklachko
Copy link

@Apokar did you try to generate Chinese music with Aiva? They have a Chinese preset, and to me it sounds pretty good (but I'm not Chinese, so...): aiva.ai

@Apokar
Copy link

Apokar commented Oct 15, 2020

@michaelklachko I'll try it ,thanks for your help !

@Randy-H0
Copy link

You can attempt to train your dataset on google colab but you'd have to restart it every time.
Your PC is also not that fast so it'll take probably more than a week.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants