Named Entity Recognition Nerkor Hubert Hungarian

🚀 Hungarian Named Entity Recognition Model with huBERT

This is a Hungarian Named Entity Recognition model based on huBERT, which can identify named entities in Hungarian text.

🚀 Quick Start

For further models, scripts and details, see our demo site.

✨ Features

• Pretrained model used: SZTAKI-HLT/hubert-base-cc
• Finetuned on NYTK-NerKor
• NE categories are: PER, LOC, MISC, ORG

🔧 Technical Details

Limitations

• max_seq_length = 128

📚 Documentation

Results

F-score: 90.18%

💻 Usage Examples

Basic Usage

from transformers import pipeline

ner = pipeline(task="ner", model="NYTK/named-entity-recognition-nerkor-hubert-hungarian")
input_text = "A Kovácsné Nagy Erzsébet nagyon jól érzi magát a Nokiánál, azonban a Németországból érkezett Kovács Péter nehezen boldogul a beilleszkedéssel."

print(ner(input_text, aggregation_strategy="simple"))

📄 License

This project is licensed under the Apache-2.0 license.

📚 Citation

If you use this model, please cite the following paper:

@inproceedings {yang-language-models,
    title = {Training language models with low resources: RoBERTa, BART and ELECTRA experimental models for Hungarian},
	booktitle = {Proceedings of 12th IEEE International Conference on Cognitive Infocommunications (CogInfoCom 2021)},
	year = {2021},
	publisher = {IEEE},
	address = {Online},
	author = {Yang, Zijian Győző and Váradi, Tamás},
	pages = {279--285}
}