Downloads · 30 days
3
2% of all-time downloads
delarosajav95/eurovoc-full-hierarchy-pipeline
eurovoc-full-hierarchy-pipeline is a text classification model from delarosajav95. Use it when you need a label for a piece of text. The card lists the license as cc-by-4.0.
- Base Model: nlpaueb/legal-bert-base-uncased - Classifier: Linear layer with 6958 EuroVoc classes - Enrichment: Real-time SPARQL queries to EUR-Lex endpoint
Downloads · 30 days
3
2% of all-time downloads
All-time downloads
150
Public
Repo size
460 MB
Likes
0
Public
Click a slice to open those files.
.bin459 MB · 100%
From the Hugging Face model README
from eurovoc_pipeline import EuroVocPipeline
# Load the pipeline
pipeline = EuroVocPipeline.from_pretrained("delarosajav95/eurovoc-full-hierarchy-pipeline")
# Predict
text = "The European Parliament adopted a resolution on climate change..."
predictions = pipeline.predict(text)
# Get formatted output
print(pipeline.format_output(predictions, format='text'))
print(pipeline.format_output(predictions, format='json'))
Training data: EUR-Lex corpus
Number of labels: 6958
Best threshold: 0.1
Top-k predictions: 3
Maximum text length: 512 tokens
Real-time hierarchy enrichment requires internet connection
First query may be slower due to SPARQL caching
If you use this model, please cite:
@software{eurovoc-full-hierarchy-pipeline,
author = {Javier de la Rosa},
title = {EuroVoc Pipeline: Automatic Concept Tagging with Hierarchy Enrichment},
year = {2024},
url = {https://huggingface.co/delarosajav95/eurovoc-full-hierarchy-pipeline}
}