Coder Social home page Coder Social logo

Demucs weights about demucs.cpp HOT 5 CLOSED

gitbearflying avatar gitbearflying commented on August 27, 2024
Demucs weights

from demucs.cpp.

Comments (5)

gitbearflying avatar gitbearflying commented on August 27, 2024 1

I answer my own question:
After downloading about 10Gb in a micromamba env, i was able to download and convert the demucs model files, the links are

https://dl.fbaipublicfiles.com/demucs/hybrid_transformer/955717e8-8726e21a.th
https://dl.fbaipublicfiles.com/demucs/hybrid_transformer/5c90dfd2-34c22ccb.th
https://dl.fbaipublicfiles.com/demucs/hybrid_transformer/f7e0c4bc-ba3fe64a.th
https://dl.fbaipublicfiles.com/demucs/hybrid_transformer/d12395a8-e57c48e6.th
https://dl.fbaipublicfiles.com/demucs/hybrid_transformer/92cfc3b6-ef3bcb9c.th
https://dl.fbaipublicfiles.com/demucs/hybrid_transformer/04573f0d-f3cf25b2.th
https://dl.fbaipublicfiles.com/demucs/hybrid_transformer/75fc33f5-1941ce65.th

To convert, these need to be placed here: ~/.cache/torch/hub/checkpoints/

To save this hassle for everybody else, i uploaded the ggml files here:
https://huggingface.co/datasets/Retrobear/demucs.cpp/tree/main

Cheers

from demucs.cpp.

gitbearflying avatar gitbearflying commented on August 27, 2024 1

Yes, after all the instructions did work. I had to remove "-e ./vendor/demucs" from requirements.txt and manually install torch and demucs. I have an old slow machine and slow internet connections, various timeouts while downloading the whole python-env etc..
Indeed then the weight downloads and conversion did work flawlessly, i put the links here just for reference as i asked for them myself. Maybe these are useful(at least to me) if another conversion to something else is needed.

from demucs.cpp.

sevagh avatar sevagh commented on August 27, 2024

Nice - did these instructions not work? https://github.com/sevagh/demucs.cpp?tab=readme-ov-file#convert-weights

I suppose they are Linux-centric (where ~/.cache/torch/hub/checkpoints is where Torch automatically downloads the weights files). That step should run when the model is initialized like so:
https://github.com/sevagh/demucs.cpp/blob/main/scripts/convert-pth-to-ggml.py#L49-L70

When demucs.pretrained.get_model is called, it should automatically download those files to Path(torch.hub.get_dir()) / "checkpoints".

from demucs.cpp.

sevagh avatar sevagh commented on August 27, 2024

Your huggingface model files might be useful to others, in case you want to add them to the README of this repo I'd be happy to accept a PR.

from demucs.cpp.

gitbearflying avatar gitbearflying commented on August 27, 2024

Your huggingface model files might be useful to others, in case you want to add them to the README of this repo I'd be happy to accept a PR.

Done :-)

from demucs.cpp.

Related Issues (12)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.