425776024/nlpcda

Health v2 · maintenance only

Adoption not part of the score

1878 stars · 171 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

nlpcda is a Python library for one-click Chinese NLP data augmentation. It provides techniques like random entity replacement, synonym substitution, homophone character swaps, character deletion, neighbor permutation, BIO-format NER augmentation, translation-based augmentation, and SimBERT-based similar sentence generation.

Use cases

• augment chinese text training data for nlp models
• generate similar sentences with bert for data augmentation
• augment NER BIO labeled data
• improve model robustness against adversarial text perturbations
• synonym replacement for chinese corpus expansion
• create more training examples from a small chinese dataset

When to choose

• you need Chinese-language text augmentation with semantic preservation
• you want multiple augmentation strategies (EDA, NER-aware, SimBERT) in one pip-installable package
• you are training BERT-style Chinese models and need more varied training data

When to avoid

• you need data augmentation for non-Chinese languages
• you expect accuracy gains in pure leaderboard-score competitions without semantic variation
• you need production text processing rather than training-data generation

Facets

library · maturity active

nlp data-generation machine-learning machine-learning data-science python cross-platform data-augmentation chinese-nlp eda bert ner text-generation simbert natural-language-processing

Member repositories

For agents

markdown · JSON · MCP: product_card(name="425776024/nlpcda")