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

app.py now supports --share #1

Open
adefossez opened this issue Jun 11, 2023 · 5 comments
Open

app.py now supports --share #1

adefossez opened this issue Jun 11, 2023 · 5 comments

Comments

@adefossez
Copy link

Hello @camenduru , thanks for the coverage and Colab on our work!
We have recently updated the app.py gradio app, so that you can pass --share to activate the shared link. this mean you can directly use the audiocraft repository, which might allow to benefit from the latest advances without having to update your fork of the repo.

@camenduru
Copy link
Owner

thanks @adefossez

@camenduru
Copy link
Owner

Hi @adefossez 👋 If we convert the video to a 3:1 aspect ratio, then people can share it on Twitter

facebookresearch/audiocraft@main...camenduru:audiocraft:dev

KubZhGozyfV8.mp4

@camenduru
Copy link
Owner

also when we switch the model, we should free the vram otherwise the vm will crash

image

@camenduru camenduru reopened this Jun 11, 2023
@promocot
Copy link

As always, no matter what the Meta developers do, everything works through the ass. you can generate a maximum of 5 songs, after which the network error and the product is not available, also pieces for more than 10 seconds are complete nonsense, nothing is clear what is playing

@Oncorporation
Copy link

Oncorporation commented Jun 19, 2023

I have recently made major improvements over at Oncorporation/audiocraft

  • I put a background image, which you can change, to make the size ratio of the video sharable. This image can include the settings used.
  • Both Melody conditioned and Fully Generated music has a 720s max length.
  • The music is now very consistent, as the settings in the UI are more specific and not limited to integers.
  • There is a hugging face version up and running for you to test Surn/UnlimitedMusicGen (Branch huggingface)
  • Volume levels are better to prevent clipping
  • waves are stereo, no longer mono
  • limiters on some variables to reduce hacking
  • lots more, but my interrupt functions need help!

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