👋 Welcome

GLiNER Community is the official Hugging Face organization for sharing and collaborating on GLiNER models. GLiNER is a Named Entity Recognition (NER) framework capable of identifying any entity type using a bidirectional transformer encoder (BERT-like). It provides a practical, efficient alternative to both traditional NER systems (limited to predefined entities) and Large Language Models (flexible but costly).

GLiNER was published at NAACL 2024 and has since grown into a thriving ecosystem of models, tools, and integrations.

🚀 Why GLiNER?

• Zero-Shot NER — Recognize any entity type without task-specific training. Just pass in your labels.
• Lightweight & Fast — Optimized to run on CPUs and consumer hardware. No GPU required.
• Fine-Tunable — Easily adapt to your domain with custom training data.
• Competitive Performance — Matches or exceeds LLMs like ChatGPT on NER tasks at a fraction of the cost.
• Multi-Task Ready — Supports NER, relation extraction, text classification, and structured data extraction.

📦 Community Models

Models hosted under this organization include the latest GLiNER releases:

Browse all GLiNER-compatible models on the Hub: Models with GLiNER library tag

⚡ Quick Start

Installation

pip install gliner -U

Usage

from gliner import GLiNER

# Load a model
model = GLiNER.from_pretrained("gliner-community/gliner_medium-v2.5", load_tokenizer=True)

# Define your text and entity labels — any labels you want!
text = """
Apple Inc. announced today that CEO Tim Cook will present the company's
new AI strategy at a press event in Cupertino on March 15, 2025.
"""

labels = ["person", "organization", "date", "location", "event"]

# Predict entities
entities = model.predict_entities(text, labels, threshold=0.5)

for entity in entities:
    print(f"{entity['text']}  =>  {entity['label']}  (score: {entity['score']:.2f})")

Output:

Apple Inc.      =>  organization  (score: 0.98)
Tim Cook        =>  person        (score: 0.97)
Cupertino       =>  location      (score: 0.95)
March 15, 2025  =>  date          (score: 0.96)

🎯 Use Cases

GLiNER powers a wide range of applications:

• Information Extraction — Pull structured data from unstructured documents
• PII Detection & Redaction — Identify sensitive information for compliance (GDPR, HIPAA, CCPA)
• Knowledge Graph Construction — Extract entities and relations to build knowledge bases
• Document Processing — Automate entity tagging in legal, financial, and medical documents
• Search & Retrieval — Enhance search pipelines with entity-aware indexing

💬 Join the Community

We'd love to have you! Connect with fellow GLiNER users and contributors:

🎮 Join our Discord →

Our Discord server is the best place to:

• Get help with GLiNER setup and usage
• Share your projects and fine-tuned models
• Discuss new features and research directions
• Connect directly with maintainers and contributors
• Stay up to date on the latest releases

Other channels

• GitHub Issues — Report bugs and request features
• GitHub Discussions — Ask questions and share ideas
• Documentation — urchade.github.io/GLiNER

📖 Citation

If you use GLiNER in your research, please cite the original paper:

@misc{zaratiana2023gliner,
    title={GLiNER: Generalist Model for Named Entity Recognition using Bidirectional Transformer},
    author={Urchade Zaratiana and Nadi Tomeh and Pierre Holat and Thierry Charnois},
    year={2023},
    eprint={2311.08526},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}

@misc{stepanov2024glinermultitaskgeneralistlightweight,
    title={GLiNER multi-task: Generalist Lightweight Model for Various Information Extraction Tasks},
    author={Ihor Stepanov and Mykhailo Shtopko},
    year={2024},
    eprint={2406.12925},
    archivePrefix={arXiv},
    primaryClass={cs.LG},
    url={https://arxiv.org/abs/2406.12925}
}

📄 License

Models in this organization are released under the Apache 2.0 License unless otherwise noted on individual model cards.

---