Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,64 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: mit
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
datasets:
|
| 4 |
+
- MultiCoNER/multiconer_v2
|
| 5 |
+
language:
|
| 6 |
+
- hi
|
| 7 |
+
metrics:
|
| 8 |
+
- f1
|
| 9 |
+
- precision
|
| 10 |
+
- recall
|
| 11 |
+
base_model:
|
| 12 |
+
- FacebookAI/xlm-roberta-large
|
| 13 |
+
pipeline_tag: token-classification
|
| 14 |
+
tags:
|
| 15 |
+
- NER
|
| 16 |
+
- Named_Entity_Recognition
|
| 17 |
+
pretty_name: MultiCoNER2 Hindi XLM-RoBERTa
|
| 18 |
+
---
|
| 19 |
+
|
| 20 |
+
**XLM-RoBERTa is fine-tuned on Hindi [MultiCoNER2](https://huggingface.co/datasets/MultiCoNER/multiconer_v2) dataset for Fine-grained Named Entity Recognition.**
|
| 21 |
+
|
| 22 |
+
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:
|
| 23 |
+
|
| 24 |
+
* Location (LOC) : Facility, OtherLOC, HumanSettlement, Station
|
| 25 |
+
* Creative Work (CW) : VisualWork, MusicalWork, WrittenWork, ArtWork, Software
|
| 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 |
+
## Model performance:
|
| 33 |
+
Precision: 76.07 <br>
|
| 34 |
+
Recall: 79.42 <br>
|
| 35 |
+
**F1: 77.71** <br>
|
| 36 |
+
|
| 37 |
+
## Training Parameters:
|
| 38 |
+
Epochs: 6 <br>
|
| 39 |
+
Optimizer: AdamW <br>
|
| 40 |
+
Learning Rate: 5e-5 <br>
|
| 41 |
+
Weight Decay: 0.01 <br>
|
| 42 |
+
Batch Size: 64 <br>
|
| 43 |
+
|
| 44 |
+
|
| 45 |
+
## Citation
|
| 46 |
+
|
| 47 |
+
If you use this model, please cite the following papers:
|
| 48 |
+
|
| 49 |
+
```bibtex
|
| 50 |
+
@inproceedings{fetahu2023multiconer,
|
| 51 |
+
title={MultiCoNER v2: a Large Multilingual dataset for Fine-grained and Noisy Named Entity Recognition},
|
| 52 |
+
author={Fetahu, Besnik and Chen, Zhiyu and Kar, Sudipta and Rokhlenko, Oleg and Malmasi, Shervin},
|
| 53 |
+
booktitle={Findings of the Association for Computational Linguistics: EMNLP 2023},
|
| 54 |
+
pages={2027--2051},
|
| 55 |
+
year={2023}
|
| 56 |
+
}
|
| 57 |
+
|
| 58 |
+
@inproceedings{kaushik2026sampurner,
|
| 59 |
+
title={SampurNER: Fine-grained Named Entity Recognition Dataset for 22 Indian Languages},
|
| 60 |
+
author={Kaushik, Prachuryya and Anand, Ashish},
|
| 61 |
+
booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
|
| 62 |
+
volume={40},
|
| 63 |
+
year={2026}
|
| 64 |
+
}
|