Add library_name and project links
Browse filesHi! I'm Niels from the Hugging Face community team.
I've opened this PR to improve the metadata and documentation for this model. Specifically:
- Added `library_name: transformers` to the metadata to enable automated code snippets and the inference widget.
- Included direct links to the AWED-FiNER research paper and GitHub repository in the model card for better discoverability.
- Formatted the existing performance and training information for better readability.
These changes help users understand the context of the model as part of the AWED-FiNER collection and make it easier to use within the Hugging Face ecosystem.
README.md
CHANGED
|
@@ -1,57 +1,65 @@
|
|
| 1 |
---
|
| 2 |
-
license: mit
|
| 3 |
-
language:
|
| 4 |
-
- brx
|
| 5 |
base_model:
|
| 6 |
- google/muril-large-cased
|
| 7 |
-
pipeline_tag: token-classification
|
| 8 |
-
tags:
|
| 9 |
-
- NER
|
| 10 |
-
- Named_Entity_Recognition
|
| 11 |
-
pretty_name: CLASSER Bodo MuRIL
|
| 12 |
datasets:
|
| 13 |
- prachuryyaIITG/CLASSER
|
|
|
|
|
|
|
|
|
|
| 14 |
metrics:
|
| 15 |
- f1
|
| 16 |
- precision
|
| 17 |
- recall
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 18 |
---
|
| 19 |
|
| 20 |
**MuRIL is fine-tuned on Bodo [CLASSER](https://huggingface.co/datasets/prachuryyaIITG/CLASSER) dataset for Fine-grained Named Entity Recognition.**
|
| 21 |
|
| 22 |
-
|
|
|
|
|
|
|
|
|
|
| 23 |
|
| 24 |
-
|
| 25 |
-
|
| 26 |
-
* Group (GRP) : MusicalGRP, PublicCORP, PrivateCORP, AerospaceManufacturer, SportsGRP, CarManufacturer, ORG
|
| 27 |
-
* Person (PER) : Scientist, Artist, Athlete, Politician, Cleric, SportsManager, OtherPER
|
| 28 |
-
* Product (PROD) : Clothing, Vehicle, Food, Drink, OtherPROD
|
| 29 |
-
* Medical (MED) : Medication/Vaccine, MedicalProcedure, AnatomicalStructure, Symptom, Disease
|
| 30 |
|
| 31 |
-
|
| 32 |
-
|
| 33 |
-
|
| 34 |
-
**
|
|
|
|
|
|
|
| 35 |
|
| 36 |
-
##
|
| 37 |
-
|
| 38 |
-
|
| 39 |
-
|
| 40 |
-
|
| 41 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 42 |
|
| 43 |
## Contributors
|
| 44 |
-
[Prachuryya Kaushik](https://www.linkedin.com/in/pkabundant/)
|
| 45 |
-
[Prof. Ashish Anand](https://www.linkedin.com/in/anandashish/)
|
| 46 |
|
| 47 |
-
CLASSER is a part of the [AWED-FiNER collection](https://huggingface.co/collections/prachuryyaIITG/awed-finer).
|
| 48 |
|
| 49 |
## Sample Usage
|
| 50 |
|
| 51 |
The AWED-FiNER agentic tool can be used to interact with expert models trained using this framework. Below is an example:
|
|
|
|
| 52 |
```bash
|
| 53 |
pip install smolagents gradio_client
|
| 54 |
```
|
|
|
|
| 55 |
```python
|
| 56 |
from tool import AWEDFiNERTool
|
| 57 |
|
|
@@ -60,8 +68,8 @@ tool = AWEDFiNERTool(
|
|
| 60 |
)
|
| 61 |
|
| 62 |
result = tool.forward(
|
| 63 |
-
text="
|
| 64 |
-
language="
|
| 65 |
)
|
| 66 |
|
| 67 |
print(result)
|
|
@@ -72,6 +80,16 @@ print(result)
|
|
| 72 |
If you use this model, please cite the following papers:
|
| 73 |
|
| 74 |
```bibtex
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 75 |
@inproceedings{kaushik-anand-2025-classer,
|
| 76 |
title = "{CLASSER}: Cross-lingual Annotation Projection enhancement through Script Similarity for Fine-grained Named Entity Recognition",
|
| 77 |
author = "Kaushik, Prachuryya and
|
|
@@ -85,29 +103,4 @@ If you use this model, please cite the following papers:
|
|
| 85 |
pages = "1745--1760",
|
| 86 |
ISBN = "979-8-89176-298-5",
|
| 87 |
}
|
| 88 |
-
|
| 89 |
-
@misc{kaushik2026awedfineragentswebapplications,
|
| 90 |
-
title={AWED-FiNER: Agents, Web applications, and Expert Detectors for Fine-grained Named Entity Recognition across 36 Languages for 6.6 Billion Speakers},
|
| 91 |
-
author={Prachuryya Kaushik and Ashish Anand},
|
| 92 |
-
year={2026},
|
| 93 |
-
eprint={2601.10161},
|
| 94 |
-
archivePrefix={arXiv},
|
| 95 |
-
primaryClass={cs.CL},
|
| 96 |
-
url={https://arxiv.org/abs/2601.10161},
|
| 97 |
-
}
|
| 98 |
-
|
| 99 |
-
@inproceedings{kaushik2026sampurner,
|
| 100 |
-
title={SampurNER: Fine-grained Named Entity Recognition Dataset for 22 Indian Languages},
|
| 101 |
-
author={Kaushik, Prachuryya and Anand, Ashish},
|
| 102 |
-
booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
|
| 103 |
-
volume={40},
|
| 104 |
-
year={2026}
|
| 105 |
-
}
|
| 106 |
-
|
| 107 |
-
@inproceedings{fetahu2023multiconer,
|
| 108 |
-
title={MultiCoNER v2: a Large Multilingual dataset for Fine-grained and Noisy Named Entity Recognition},
|
| 109 |
-
author={Fetahu, Besnik and Chen, Zhiyu and Kar, Sudipta and Rokhlenko, Oleg and Malmasi, Shervin},
|
| 110 |
-
booktitle={Findings of the Association for Computational Linguistics: EMNLP 2023},
|
| 111 |
-
pages={2027--2051},
|
| 112 |
-
year={2023}
|
| 113 |
-
}
|
|
|
|
| 1 |
---
|
|
|
|
|
|
|
|
|
|
| 2 |
base_model:
|
| 3 |
- google/muril-large-cased
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
datasets:
|
| 5 |
- prachuryyaIITG/CLASSER
|
| 6 |
+
language:
|
| 7 |
+
- brx
|
| 8 |
+
license: mit
|
| 9 |
metrics:
|
| 10 |
- f1
|
| 11 |
- precision
|
| 12 |
- recall
|
| 13 |
+
pipeline_tag: token-classification
|
| 14 |
+
library_name: transformers
|
| 15 |
+
tags:
|
| 16 |
+
- NER
|
| 17 |
+
- Named_Entity_Recognition
|
| 18 |
+
pretty_name: CLASSER Bodo MuRIL
|
| 19 |
---
|
| 20 |
|
| 21 |
**MuRIL is fine-tuned on Bodo [CLASSER](https://huggingface.co/datasets/prachuryyaIITG/CLASSER) dataset for Fine-grained Named Entity Recognition.**
|
| 22 |
|
| 23 |
+
This model is part of the **AWED-FiNER** collection, as presented in the paper [AWED-FiNER: Agents, Web applications, and Expert Detectors for Fine-grained Named Entity Recognition across 36 Languages for 6.6 Billion Speakers](https://huggingface.co/papers/2601.10161).
|
| 24 |
+
|
| 25 |
+
- **GitHub Repository:** [AWED-FiNER](https://github.com/PrachuryyaKaushik/AWED-FiNER)
|
| 26 |
+
- **Interactive Demo:** [AWED-FiNER Space](https://huggingface.co/spaces/prachuryyaIITG/AWED-FiNER)
|
| 27 |
|
| 28 |
+
### Tagset Mapping
|
| 29 |
+
The model uses the fine-grained tagset from [MultiCoNER2](https://huggingface.co/datasets/MultiCoNER/multiconer_v2). The mapping from fine to coarse level tags is as follows:
|
|
|
|
|
|
|
|
|
|
|
|
|
| 30 |
|
| 31 |
+
* **Location (LOC)** : Facility, OtherLOC, HumanSettlement, Station
|
| 32 |
+
* **Creative Work (CW)** : VisualWork, MusicalWork, WrittenWork, ArtWork, Software
|
| 33 |
+
* **Group (GRP)** : MusicalGRP, PublicCORP, PrivateCORP, AerospaceManufacturer, SportsGRP, CarManufacturer, ORG
|
| 34 |
+
* **Person (PER)** : Scientist, Artist, Athlete, Politician, Cleric, SportsManager, OtherPER
|
| 35 |
+
* **Product (PROD)** : Clothing, Vehicle, Food, Drink, OtherPROD
|
| 36 |
+
* **Medical (MED)** : Medication/Vaccine, MedicalProcedure, AnatomicalStructure, Symptom, Disease
|
| 37 |
|
| 38 |
+
## Model Performance
|
| 39 |
+
- **Precision:** 73.83
|
| 40 |
+
- **Recall:** 76.37
|
| 41 |
+
- **F1 Score:** 75.08
|
| 42 |
+
|
| 43 |
+
## Training Parameters
|
| 44 |
+
- **Epochs:** 6
|
| 45 |
+
- **Optimizer:** AdamW
|
| 46 |
+
- **Learning Rate:** 5e-5
|
| 47 |
+
- **Weight Decay:** 0.01
|
| 48 |
+
- **Batch Size:** 64
|
| 49 |
|
| 50 |
## Contributors
|
| 51 |
+
[Prachuryya Kaushik](https://www.linkedin.com/in/pkabundant/) and [Prof. Ashish Anand](https://www.linkedin.com/in/anandashish/).
|
|
|
|
| 52 |
|
| 53 |
+
CLASSER is a part of the [AWED-FiNER collection](https://huggingface.co/collections/prachuryyaIITG/awed-finer).
|
| 54 |
|
| 55 |
## Sample Usage
|
| 56 |
|
| 57 |
The AWED-FiNER agentic tool can be used to interact with expert models trained using this framework. Below is an example:
|
| 58 |
+
|
| 59 |
```bash
|
| 60 |
pip install smolagents gradio_client
|
| 61 |
```
|
| 62 |
+
|
| 63 |
```python
|
| 64 |
from tool import AWEDFiNERTool
|
| 65 |
|
|
|
|
| 68 |
)
|
| 69 |
|
| 70 |
result = tool.forward(
|
| 71 |
+
text="अमिताभ बच्चनआ सासे मुंदांखा फावखुंगुर।",
|
| 72 |
+
language="Bodo"
|
| 73 |
)
|
| 74 |
|
| 75 |
print(result)
|
|
|
|
| 80 |
If you use this model, please cite the following papers:
|
| 81 |
|
| 82 |
```bibtex
|
| 83 |
+
@misc{kaushik2026awedfineragentswebapplications,
|
| 84 |
+
title={AWED-FiNER: Agents, Web applications, and Expert Detectors for Fine-grained Named Entity Recognition across 36 Languages for 6.6 Billion Speakers},
|
| 85 |
+
author={Prachuryya Kaushik and Ashish Anand},
|
| 86 |
+
year={2026},
|
| 87 |
+
eprint={2601.10161},
|
| 88 |
+
archivePrefix={arXiv},
|
| 89 |
+
primaryClass={cs.CL},
|
| 90 |
+
url={https://arxiv.org/abs/2601.10161},
|
| 91 |
+
}
|
| 92 |
+
|
| 93 |
@inproceedings{kaushik-anand-2025-classer,
|
| 94 |
title = "{CLASSER}: Cross-lingual Annotation Projection enhancement through Script Similarity for Fine-grained Named Entity Recognition",
|
| 95 |
author = "Kaushik, Prachuryya and
|
|
|
|
| 103 |
pages = "1745--1760",
|
| 104 |
ISBN = "979-8-89176-298-5",
|
| 105 |
}
|
| 106 |
+
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|