search

IraKorshunova / folk-rnn

folk music modelling with LSTM
MIT License
341 stars 68 forks source link

Package #7

Closed tobyspark closed 2 years ago

tobyspark commented 6 years ago

This adds packaging information to folk_rnn, and extracts the core functionality of folk_rnn into a class. This is in aid of creating a folk_rnn task as part of a web-app, which will require installation on the server and a programmatic interface to the folk_rnn functionality.

e.g. https://github.com/tobyspark/folk-rnn-webapp/commit/d60f73a6baa30b5c4eca329b37a4b52c9ecc9936

The packaging is not exhaustive. It's certainly not enough to upload to PyPi, which would be desirable to ensure version control on import to production servers. This is the minimum viable to be able to pip install -e . on a clone of the repo, e.g. https://github.com/tobyspark/folk-rnn-webapp/commit/6db57e2095342c2f93852c35d0d83f22bac71b92

This refactored codebase appears to produce identical output – see note in https://github.com/tobyspark/folk-rnn/commit/381184a2d6659a47520cedd6d4dfa7bb1c5189f7 – but no guarantees are made! Please satisfy yourselves before merging.

IraKorshunova commented 6 years ago

Hi Toby,

I'm very sorry, but I haven't worked on this project for ages, so it wouldn't be easy for me to check your PR and make sure that everything is still working. I'll keep this PR open, so people can find your fork

tobyspark commented 6 years ago

Ira – Sure. I'm working with @boblsturm who has been updating the readme with the research output. Perhaps he'll drop a note there when the envisaged folk-rnn community site goes live, linking to the folk-rnn code (or whatever future research) driving it.