Update README.md
Browse files
README.md
CHANGED
|
@@ -7,6 +7,7 @@ tags:
|
|
| 7 |
- named-entity-recognition
|
| 8 |
- bert
|
| 9 |
- token-classification
|
|
|
|
| 10 |
datasets:
|
| 11 |
- custom
|
| 12 |
metrics:
|
|
@@ -22,15 +23,27 @@ widget:
|
|
| 22 |
|
| 23 |
## Model Description
|
| 24 |
|
| 25 |
-
This is an Arabic Named Entity Recognition (NER) model fine-tuned on BERT architecture specifically for Arabic text processing. The model is based on `MutazYoune/ARAB_BERT` and
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 26 |
|
| 27 |
## Model Details
|
| 28 |
|
| 29 |
- **Model Type:** Token Classification (NER)
|
| 30 |
- **Language:** Arabic (ar)
|
| 31 |
- **Base Model:** MutazYoune/ARAB_BERT
|
| 32 |
-
- **Dataset:** augmented_pattern2
|
| 33 |
-
- **Task:** Named Entity Recognition
|
| 34 |
|
| 35 |
## Training Configuration
|
| 36 |
|
|
@@ -65,26 +78,3 @@ ner_pipeline = pipeline("ner",
|
|
| 65 |
text = "أحمد محمد يعمل في شركة جوجل في الرياض"
|
| 66 |
entities = ner_pipeline(text)
|
| 67 |
print(entities)
|
| 68 |
-
```
|
| 69 |
-
|
| 70 |
-
## Model Performance
|
| 71 |
-
|
| 72 |
-
This model was trained on the complete dataset without validation split for final production use.
|
| 73 |
-
|
| 74 |
-
## Training Data
|
| 75 |
-
|
| 76 |
-
The model was trained on custom Arabic NER dataset:
|
| 77 |
-
- Dataset type: augmented_pattern2
|
| 78 |
-
- Combined training and test data for final model
|
| 79 |
-
|
| 80 |
-
## Citation
|
| 81 |
-
|
| 82 |
-
```bibtex
|
| 83 |
-
@misc{arabic-ner-bert,
|
| 84 |
-
title={Arabic BERT NER Model},
|
| 85 |
-
author={Trained on Kaggle},
|
| 86 |
-
year={2025},
|
| 87 |
-
publisher={Hugging Face},
|
| 88 |
-
url={https://huggingface.co/MutazYoune/Arabic-NER-PII}
|
| 89 |
-
}
|
| 90 |
-
```
|
|
|
|
| 7 |
- named-entity-recognition
|
| 8 |
- bert
|
| 9 |
- token-classification
|
| 10 |
+
- pii
|
| 11 |
datasets:
|
| 12 |
- custom
|
| 13 |
metrics:
|
|
|
|
| 23 |
|
| 24 |
## Model Description
|
| 25 |
|
| 26 |
+
This is an Arabic Named Entity Recognition (NER) model fine-tuned on BERT architecture specifically for Arabic text processing and PII detection. The model is based on `MutazYoune/ARAB_BERT` and trained to identify and mask Personally Identifiable Information (PII) in Arabic sentences.
|
| 27 |
+
|
| 28 |
+
PII categories covered include:
|
| 29 |
+
|
| 30 |
+
- Personal names (first, middle, family)
|
| 31 |
+
- Phone numbers
|
| 32 |
+
- Email addresses
|
| 33 |
+
- Physical addresses
|
| 34 |
+
- National ID numbers
|
| 35 |
+
- Bank account information
|
| 36 |
+
- Dates of birth
|
| 37 |
+
|
| 38 |
+
The model was developed and submitted as part of the **Arabic PII Redaction Challenge** hosted on Hugging Face.
|
| 39 |
|
| 40 |
## Model Details
|
| 41 |
|
| 42 |
- **Model Type:** Token Classification (NER)
|
| 43 |
- **Language:** Arabic (ar)
|
| 44 |
- **Base Model:** MutazYoune/ARAB_BERT
|
| 45 |
+
- **Dataset:** augmented_pattern2 (custom Arabic PII NER dataset)
|
| 46 |
+
- **Task:** Named Entity Recognition and PII redaction
|
| 47 |
|
| 48 |
## Training Configuration
|
| 49 |
|
|
|
|
| 78 |
text = "أحمد محمد يعمل في شركة جوجل في الرياض"
|
| 79 |
entities = ner_pipeline(text)
|
| 80 |
print(entities)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|