---
language:
- brx
- lus
- mni
license: mit
size_categories:
- 10M<n<100M
task_categories:
- token-classification
- other
pretty_name: FiNERVINER
---

# FiNERVINER: Fine-grained Named Entity Recognition for Vulnerable languages of India's North Eastern Region

**FiNERVINER** is a high-quality fine-grained named entity recognition dataset created through annotation projection method. 

The vulnerable languages are: Bodo (brx), Mizo (lus), and Manipuri (mni).

## FiNERVINER Dataset Statistics

<table>
<thead>
    <tr>
      <th rowspan="2">Language</th>
      <th colspan="3">Train set</th>
      <th colspan="3">Development set</th>
      <th colspan="4">Test set</th>
    </tr>
    <tr>
      <th>Sentences</th><th>Entities</th><th>Tokens</th>
      <th>Sentences</th><th>Entities</th><th>Tokens</th>
      <th>Sentences</th><th>Entities</th><th>Tokens</th><th>IAA&nbsp;(κ)</th>
    </tr>
  </thead>
  <tbody align="center">
    <tr>
      <td><b>Bodo (brx)</b></td>
      <td>212,835</td><td>302,713</td><td>2,958,455</td>
      <td>23,649</td><td>33,808</td><td>329,145</td>
      <td>1,000</td><td>1,423</td><td>14,082</td><td><b>0.875</b></td>
    </tr>
    <tr>
      <td><b>Mizo (lus)</b></td>
      <td>177,224</td><td>252,767</td><td>2,515,386</td>
      <td>19,692</td><td>28,143</td><td>279,681</td>
      <td>1,000</td><td>1,384</td><td>14,330</td><td><b>0.811</b></td>
    </tr>
    <tr>
      <td><b>Manipuri (mni)</b></td>
      <td>239,813</td><td>302,713</td><td>4,422,373</td>
      <td>26,646</td><td>38,330</td><td>484,212</td>
      <td>1,000</td><td>1,426</td><td>18,765</td><td><b>0.821</b></td>
    </tr>
  </tbody>
</table>

*Note: IAA (Inter-Annotator Agreement) scores are represented using Cohen's κ.*

## Contributors
[Prachuryya Kaushik](https://www.linkedin.com/in/pkabundant/) <br>
[Prof. Ashish Anand](https://www.linkedin.com/in/anandashish/)


FiNERVINER is a part of the [AWED-FiNER](https://github.com/PrachuryyaKaushik/AWED-FiNER) ecosystem: [**Paper**](https://huggingface.co/papers/2601.10161) | [**GitHub**](https://github.com/PrachuryyaKaushik/AWED-FiNER) | [**Interactive Demo**](https://huggingface.co/spaces/prachuryyaIITG/AWED-FiNER)

## Sample Usage

You can use the AWED-FiNER agentic tool to interact with expert models trained using this framework. Below is an example using the `smolagents` library:

```python
from smolagents import CodeAgent, HfApiModel
from tool import AWEDFiNERTool

# Initialize the expert tool
ner_tool = AWEDFiNERTool()

# Initialize the agent (using a model of your choice)
agent = CodeAgent(tools=[ner_tool], model=HfApiModel())

# The agent will automatically use AWED-FiNER for specialized NER
# Case: Processing a vulnerable language (Bodo)
agent.run("Recognize the named entities in this Bodo sentence: 'बिथाङा दिल्लियाव थाङो।'")
```

## Citation

If you use this dataset, please cite the following papers:

```bibtex
@inproceedings{kaushik2026finerviner,
  title={FiNERVINER: Fine-grained Named Entity Recognition for Vulnerable languages of India's North Eastern Region},
  author={Kaushik, Prachuryya and Anand, Ashish},
  booktitle={Proceedings of the Fifteenth Language Resources and Evaluation Conference},
  volume={15},
  year={2026}
}

@inproceedings{kaushik2026sampurner,
  title={SampurNER: Fine-grained Named Entity Recognition dataset for 22 Indian Languages},
  author={Kaushik, Prachuryya and Anand, Ashish},
  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
  volume={40},
  year={2026}
}

@misc{kaushik2026awedfineragentswebapplications,
      title={AWED-FiNER: Agents, Web applications, and Expert Detectors for Fine-grained Named Entity Recognition across 36 Languages for 6.6 Billion Speakers}, 
      author={Prachuryya Kaushik and Ashish Anand},
      year={2026},
      eprint={2601.10161},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2601.10161}, 
}
```