#196 · Primary category： NLP Tools & Text Processing
rebel
REBEL is a seq2seq model that simplifies Relation Extraction (EMNLP 2021).
Project last updated：11/09/23
GitHub Stars
574
Forks
76
Contributors
4
License
Other
Why we included this project
REBEL extracts structured triplets from raw text in a single pass, which is exactly what you need when assembling a knowledge graph or a fact-checking pipeline. Rather than chaining a separate entity recognizer to a relation classifier, it frames the whole job as text-to-text generation on top of BART: give it "Punta Cana is a resort town in the Dominican Republic" and it returns explicit head-relation-tail triples covering more than 200 relation types. The pretrained model loads through Hugging Face with a few lines of transformers code, and the mREBEL variant does the same work in languages beyond English. A Streamlit demo and a spaCy plugin let you try it on your own documents before committing to a larger pipeline. It is a research-grade baseline that is easy to run.
Articles for this project
No articles for this project yet.
To suggest a topic or contribute an article, contact us.
Related projects in this category
flair
A very simple framework for state-of-the-art Natural Language Processing (NLP)
compromise
modest natural-language processing
tokenizers
💥 Fast State-of-the-Art Tokenizers optimized for Research and Production
CoreNLP
CoreNLP: A Java suite of core NLP tools for tokenization, sentence segmentation, NER, parsing, coreference, sentiment analysis, etc.
Chinese-Word-Vectors
100+ Chinese Word Vectors 上百种预训练中文词向量