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

2.0.4 - 2.1.1 broken on Wine/Linux, 2.0.3 works #80

Open
infinity0 opened this issue Nov 26, 2024 · 4 comments
Open

2.0.4 - 2.1.1 broken on Wine/Linux, 2.0.3 works #80

infinity0 opened this issue Nov 26, 2024 · 4 comments

Comments

@infinity0
Copy link

All songs fail to demux even when I set CPU mode on. Console logs contains:

Trace: [Error: ENOENT: no such file or directory, scandir 'C:\users\infinity0\Temp\StemRoller-ohuoIY\separated\htdemucs_ft'] {
  errno: -4058,
  code: 'ENOENT',
  syscall: 'scandir',
  path: 'C:\\users\\infinity0\\Temp\\StemRoller-ohuoIY\\separated\\htdemucs_ft'
}
    at processVideo (G:\bin\stemroller-2.0.4-win-cuda\resources\app.asar\main-src\processQueue.cjs:344:13)

So, looks like something is wrong with the tempdir handling.

@iffyloop
Copy link
Member

FYI: it will probably be easier to install Node.js/NPM and Python manually via your package manager, install Demucs globally via pip, then run StemRoller in development mode, instead of using Wine. I'm not currently maintaining pre-built binaries for Linux but there's no Windows-specific dependencies that would prevent StemRoller from running on Linux.

@infinity0
Copy link
Author

Good to know, thanks! I'll give it a go. Do you know if ROCm (amdgpu) support is feasible and would you take a PR for it?

@iffyloop
Copy link
Member

iffyloop commented Nov 27, 2024

Yes - it should work pretty much automatically if you install a version of PyTorch which is compatible with ROCm! If you run into issues, you can try passing additional arguments into the Demucs process here: https://github.com/stemrollerapp/stemroller/blob/main/main-src/processQueue.cjs#L302 (I wonder if it needs extra arguments in case it doesn't select the correct GPU? I don't have any CUDA or ROCm device unfortunately, so can't test myself...)

Thanks for offering to create a PR - please let me know if it takes any code changes to get ROCm working, and let's discuss further after that.

@iffyloop
Copy link
Member

(Select Linux and ROCm at https://pytorch.org/get-started/locally/ to find out how to install PyTorch correctly)

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

2 participants