Skip to content

Machine Learning pipeline and API for the United Nations Digital Library

Notifications You must be signed in to change notification settings

ClementSicard/un-ml-pipeline

Repository files navigation

Machine Learning pipeline for the United Nations Digital Library 🇺🇳

Style Packages Linter VersionBuilt with HuggingFace Built with spaCy

Pipeline description

Diagram

Summarization models

Model File Paper Authors Year HuggingFace 🤗 model
DistillBART-CNN DistillBARTCNN.py arXiv link Shleifer et al. 2020 Link
DistillBART-XSUM DistillBARTXSUM.py arXiv link Shleifer et al. 2020 Link
DistilPegasusCNN DistilPegasusCNN.py arXiv link Shleifer et al. 2020 Link
Longformer LED.py arXiv link Beltagy et al. 2020 Link
LongT5 LongT5.py arXiv link Guo et al. 2022 Link

NER models

Model File Paper Authors Year HuggingFace 🤗 model
RoBERTa RoBERTa.py arXiv link Liu et al. 2019 Link
FLERT FLERT.py arXiv link Akbik et al. 2020 Link
spaCyNER spaCy.py - spaCy 2023 Link (on spaCy website)

Usage

TODO