---
license: mit
language:
- as
metrics:
- f1
- precision
- recall
base_model:
- google/muril-large-cased
pipeline_tag: token-classification
tags:
- NER
- Named_Entity_Recognition
pretty_name: APTFiNER Tamil MuRIL
datasets:
- prachuryyaIITG/APTFiNER
---

**MuRIL is fine-tuned on Assamese APTFiNER dataset for Fine-grained Named Entity Recognition.**

The tagset of [MultiCoNER2](https://huggingface.co/datasets/MultiCoNER/multiconer_v2) is a fine-grained tagset. The fine to coarse level mapping of the tags are as follows:

  * Location (LOC) : Facility, OtherLOC, HumanSettlement, Station
  * Creative Work (CW) : VisualWork, MusicalWork, WrittenWork, ArtWork, Software
  * Group (GRP) : MusicalGRP, PublicCORP, PrivateCORP, AerospaceManufacturer, SportsGRP, CarManufacturer, ORG
  * Person (PER) : Scientist, Artist, Athlete, Politician, Cleric, SportsManager, OtherPER
  * Product (PROD) : Clothing, Vehicle, Food, Drink, OtherPROD
  * Medical (MED) : Medication/Vaccine, MedicalProcedure, AnatomicalStructure, Symptom, Disease

## Model performance:
Precision: 62.62 <br>
Recall: 67.98 <br>
**F1: 65.19** <br>

## Training Parameters:
Epochs: 6 <br>
Optimizer: AdamW <br>
Learning Rate: 5e-5 <br>
Weight Decay: 0.01 <br>
Batch Size: 64 <br>

## Contributors
[Prachuryya Kaushik](https://www.linkedin.com/in/pkabundant/) <br>
[Adittya Gupta](https://www.linkedin.com/in/adittya-gupta-b64356224/) <br>
[Ajanta Maurya](https://www.linkedin.com/in/ajanta-maurya/) <br>
[Gautam Sharma](https://www.linkedin.com/in/g-s01/) <br>
[Prof. V Vijaya Saradhi](https://www.linkedin.com/in/vijaya-saradhi-a90a604/) <br>
[Prof. Ashish Anand](https://www.linkedin.com/in/anandashish/)

APTFiNER 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 model, please cite the following papers:

```bibtex
@inproceedings{kaushik2026aptfiner,
  title={APTFiNER: Annotation Preserving Translation for Fine-grained Named Entity Recognition},
  author={Kaushik, Prachuryya and Gupta, Adittya and Maurya, Ajanta and Sharma, Gautam and Saradhi, Vijaya V 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}
}

@inproceedings{fetahu2023multiconer,
  title={MultiCoNER v2: a Large Multilingual dataset for Fine-grained and Noisy Named Entity Recognition},
  author={Fetahu, Besnik and Chen, Zhiyu and Kar, Sudipta and Rokhlenko, Oleg and Malmasi, Shervin},
  booktitle={Findings of the Association for Computational Linguistics: EMNLP 2023},
  pages={2027--2051},
  year={2023}
}
```