QNLP – Full Kazakh NLP Suite

QNLP is a comprehensive and modular library for automatic processing of the Kazakh language. It combines modern transformer-based models, classic rule-based morphological processing, and production-ready deployment tools (API, CLI, ONNX).

Features

KazBERT + BiLSTM + CRF: High-accuracy POS and NER tagging
QA Model: SQuAD-style question answering
SRL: Semantic role labeling using transformer models
Coreference Resolution: Rule-based and model-enhanced
Morphology: Lemmatizer, generator, morpheme analyzer
Tokenizer: BPE, SentencePiece support
ONNX: Export models for production
API: Flask + Swagger + CLI support
Augmentation: NER-specific synthetic data generator
Unit Testing: All modules with test coverage

Models

Model	Architecture	Description
`kazner-crf`	KazBERT + BiLSTM + CRF	Named Entity Recognition
`kazpos-crf`	KazBERT + BiLSTM + CRF	Part-of-Speech Tagging
`kazqa`	KazBERT + QA Head	Question Answering
`kazsrl`	KazBERT + Classifier	Semantic Role Labeling
`kazcoref`	Rule-based + Span Linking	Coreference Resolution

Directory Structure

qaznlp/
├── models/
├── morphology/
├── tokenizer/
├── training/
├── augment/
├── server/
├── scripts/
├── data/
├── tests/
└── notebooks/

Demo & Testing

test_all.py: Run all tests
app.py: REST API with Swagger UI
export_onnx.py: Export all models to ONNX

License

For academic use, cite as: QNLP: An Open-Source Suite for Kazakh Language Processing (2025)

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support