yueyu1030/COSINE

[NAACL 2021] This is the code for our paper `Fine-Tuning Pre-trained Language Model with Weak Supervision: A Contrastive-Regularized Self-Training Approach'.

42
/ 100
Emerging

This project helps data scientists and machine learning engineers build powerful text classification models even when they only have access to weakly labeled data. It takes your text data and a set of weak labels (which might be noisy or incomplete) and outputs a highly accurate text classification model. This is for professionals who need to categorize large volumes of text but lack the resources for extensive manual data labeling.

206 stars. No commits in the last 6 months.

Use this if you need to train a robust text classification, relation extraction, or word sense disambiguation model but only have weakly labeled datasets or noisy annotations.

Not ideal if you already have a large, high-quality, fully supervised dataset for your text classification task.

text-classification natural-language-processing weak-supervision information-extraction sentiment-analysis
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 16 / 25

How are scores calculated?

Stars

206

Forks

26

Language

Python

License

MIT

Last pushed

Aug 17, 2022

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/nlp/yueyu1030/COSINE"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.