Metadata-Version: 2.1
Name: mderanklib
Version: 0.1.0
Summary: Librería para realizar ranking multilingüe de entidades.
Home-page: https://github.com/oeg-upm/mderanklib
Author: Pablo Calleja
Author-email: pablo@gmail.com
License: Apache
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: pandas==1.3.5
Requires-Dist: StanfordCoreNLP==3.9.1.1
Requires-Dist: transformers==4.37.2
Requires-Dist: accelerate==0.5.1
Requires-Dist: pke==2.0.0
Requires-Dist: torch==2.1.2
Requires-Dist: torchvision==0.16.2
Requires-Dist: torchaudio==2.1.2
Requires-Dist: nltk==3.8.1
Requires-Dist: allennlp==3.1.0

# mderanklib
Library for MDERank model to AKE

It has been adapted and improved to work also in Spanish language. 

Original Paper: MDERank: A Masked Document Embedding Rank Approach for Unsupervised Keyphrase Extraction
Original Repo: https://github.com/LinhanZ/mderank



## Install

Use requirements.txt to install the required libraries. You can use torch with and without gpu

Download also the required spacy models of Spanish/English

```
python -m spacy download en_core_web_sm
python -m spacy download es_core_news_sm

```



## Run 

execute run.sh file




## Acknowledges 







