AnnotationVocabulary

This is the official repository of the paper Annotation Vocabulary (Might Be) All You Need.

The data and models were developed using a proprietary internal developers package (udev), for which we've refactored a public version (udev_public). This repo is set up with the code intact, but currently, you will need to modify into your own files / file structure. For example, if you want to train a camp model you will have to place the udev_public file in a location that the train.py can see, or use import sys to make the parent folder available.

Models, training data

Huggingface

Downstream protein tasks

Huggingface

Demo

gsm_demo_github.mp4

If you use any of our models, data, or code, please cite the following paper (and be aware of the GPL-3.0 license)

@article{hallee2024annotation,
      title={Annotation Vocabulary(Might Be) All You Need}, 
      author={Logan Hallee and Niko Rafailidis and Colin Horger and David Hong and Jason P. Gleghorn},
      year={2024},
      eprint={},
      archivePrefix={biorXiv},
      primaryClass={cs.LG}
}