ClaudiaShu/UNA
This is the official code of our Paper "Unsupervised hard Negative Augmentation for contrastive learning"
This project helps machine learning engineers and researchers improve their natural language processing models. It takes a training dataset of text and generates additional 'hard negative' examples using TF-IDF, then outputs an enhanced dataset for training. This process helps create more robust and accurate sentence embeddings for tasks like semantic similarity.
No commits in the last 6 months.
Use this if you are a machine learning engineer working on natural language processing and want to improve the robustness of your sentence embedding models by augmenting your training data with challenging negative examples.
Not ideal if you are looking for a pre-trained model to use directly, as this project focuses on data augmentation for model training.
Stars
8
Forks
—
Language
Python
License
—
Category
Last pushed
Jan 18, 2024
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/nlp/ClaudiaShu/UNA"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
dsfsi/textaugment
TextAugment: Text Augmentation Library
425776024/nlpcda
一键中文数据增强包 ； NLP数据增强、bert数据增强、EDA：pip install nlpcda
google-research/uda
Unsupervised Data Augmentation (UDA)
searchableai/KitanaQA
KitanaQA: Adversarial training and data augmentation for neural question-answering models
SanghunYun/UDA_pytorch
UDA(Unsupervised Data Augmentation) implemented by pytorch