Code Monkey home page Code Monkey logo

Comments (2)

jongwook avatar jongwook commented on August 13, 2024

PyPI requires the upload size to be <60 MB, but we have ~80 MB. Options are:

  1. Retrain the model with slightly lower model capacity multiplier. The current multiplier is 32, and something like 24 or 28 would work
  2. Keep the model capacity, but alter the float32 weights so that it only have about 24 bits of entropy, so that when compressed, the file size is smaller than 60MB.
  3. Host the model file somewhere else, like Github, and make the setup.py script to download it.
  4. Submit the ticket to the warehouse for exception.

I did reserve the name crepe on PyPI.

from crepe.

justinsalamon avatar justinsalamon commented on August 13, 2024

I'm not a fan of (1), because we should aim to release a model that's identical (in architecture/capacity) to the one reported in the paper. (2) might work but I'd be worried about fiddling with the weights post-hoc without knowing whether/how that impacts model performance. (3) would be the cleanest in terms of keeping the model untouched, but might be problematic because we're decoupling the model file from the package version, which might cause problems if we upload updated models in the future. (4) In principle I see no reason not to try this, but I'm guessing it would take a while and there's no guarantee it'll be approved.

TL;DR - not sure what the best option is here...

from crepe.

Related Issues (20)

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.