# Yambeta NER Model
## Model Description
This model was trained to perform Named Entity Recognition (NER) on Yambeta text using the BERT architecture. It recognizes entities such as persons (PER), locations (LOC), and organizations (ORG) in Yambeta, a Bantu language spoken in Cameroon.

- **Model Type**: BERT-based token classification model.
- **Developed by**: DS4H-ICTU Research Group.
- **Language**: Yambeta (Bantu language from Cameroon).
- **License**: Apache 2.0 (or specify if different).

## Training Data
The model was fine-tuned on a dataset consisting of Yambeta Bible text, annotated for NER tasks. The dataset contains entities categorized as persons, locations, and organizations.

## Usage
You can use this model for NER tasks in Yambeta text by loading it through the Hugging Face `transformers` library:
```python
from transformers import AutoTokenizer, AutoModelForTokenClassification

tokenizer = AutoTokenizer.from_pretrained("DS4H-ICTU/yat-bert-tokenizer")
model = AutoModelForTokenClassification.from_pretrained("DS4H-ICTU/yat-ner-model")

## Evaluation
The model achieves high precision, recall, and F1 scores on the validation set, demonstrating strong performance for recognizing named entities in Yambeta text.

## Bias, Risks, and Limitations
Biases: The training data is sourced from religious texts, which may introduce biases specific to this domain. The model may not generalize well to non-religious texts in Yambeta.
Out-of-Scope Use: The model is designed for Yambeta NER tasks and may not perform well for other languages or applications outside NER.

## Citation
If you use this model in your work, please cite it as follows:

@misc{yambeta_ner_model,
  title = {Yambeta NER Model},
  author = {Dr.-Ing. Philippe Tamla},
  year = {2024},
  publisher = {Hugging Face},
  url = {https://huggingface.co/DS4H-ICTU/yat-ner-model}
}

## Contact Information
For more information, contact the developers at: philiptamla@gmail.com