-----------------------------------------------------------------
README - Domain Labeling

Sapienza NLP Group
Sapienza University of Rome
http://nlp.uniroma1.it
-----------------------------------------------------------------

This package contains a software whose function is to 
automatically tag dictionary glosses with domains of knowledge 
labels. 

The software exploits the representational power of 
multilingual BERT (M-BERT) and makes use of a simple yet 
efficient Sequence Classification architecture, composed of a 
multilingual BERT encoder -- which encodes the input -- and a 
linear layer, to perform the classification.

--------
CONTENTS
--------

This package contains the following components:

├── config
│   └── custom.jsonnet   # base jsonnet configuration file
├── data                 # folder for data
├── LICENSE
├── models
│   ├── released         # folder for released models
│   └── trained          # folder for trained models
├── src                  # source code
│   ├── allen_elements   # allennlp-related code
│   ├── serve.py         # simple cli interactive demo
│   └── main.py          # training entry point
├── requirements.txt
├── README
└── USAGE.md

------------
REQUIREMENTS
------------

A python 3.7 installation (possibly in an environment manager).

------------
INSTALLATION
------------

Before running the following command, it is advised to create a new environment
(e.g., with conda) to avoid conflicts with your current one.

Install requirements via pip install -r requirements.txt.

The code is based off of AllenNLP's library. See usage instructions at USAGE.md.

-------
AUTHORS
-------

Federico Martelli, Sapienza University of Rome
([email protected])

Roberto Navigli, Sapienza University of Rome
([email protected])

Acknowledgments go to Niccolò Campolungo, Babelscape
([email protected]), for his contribution to the project.

---------
COPYRIGHT
---------

This software is licensed under a Creative Commons Attribution-Noncommercial-
Share Alike 4.0 License. See the LICENSE file for details.

---------------
ACKNOWLEDGMENTS
---------------

This software is an output of the ELEXIS project (https://elex.is). This project has
received funding from the European Union's Horizon 2020 research and innovation
programme under grant agreement No. 731015.