|
|
--- |
|
|
license: apache-2.0 |
|
|
datasets: |
|
|
- risqaliyevds/uzbek_ner |
|
|
language: |
|
|
- uz |
|
|
metrics: |
|
|
- precision |
|
|
- f1 |
|
|
- recall |
|
|
- accuracy |
|
|
base_model: |
|
|
- FacebookAI/xlm-roberta-base |
|
|
pipeline_tag: token-classification |
|
|
--- |
|
|
|
|
|
# NER Model for Uzbek Language (XLM-RoBERTa-based) |
|
|
|
|
|
This is a Named Entity Recognition (NER) model trained for the Uzbek language based on the XLM-RoBERTa architecture. It is fine-tuned to classify entities into categories such as location, person, organization, and other types. |
|
|
|
|
|
## Model Details |
|
|
|
|
|
- **Model Type**: XLM-RoBERTa (Transformer-based) |
|
|
- **Task**: Named Entity Recognition (NER) |
|
|
- **Training Data**: Custom dataset with labeled named entities for Uzbek language. |
|
|
- **Categories**: |
|
|
- `B-LOC` (Location) |
|
|
- `B-PERSON` (Person) |
|
|
- `B-ORG` (Organization) |
|
|
- `B-PRODUCT` (Product) |
|
|
- `B-DATE` (Date) |
|
|
- `B-TIME` |
|
|
- `B-LANGUAGE` |
|
|
- `B-GPE` |
|
|
|
|
|
## Metrics |
|
|
- **Validation accuracy = 0.9793** |
|
|
- **val_loss: 0.1141** |
|
|
|
|
|
- **Precision: 0.97** |
|
|
- **Recall: 0.97** |
|
|
- **F1-Score: 0.97** |
|
|
|
|
|
## Usage |
|
|
|
|
|
You can use this model with the Hugging Face Transformers library to perform NER tasks on your own Uzbek language text. |
|
|
|
|
|
```python |
|
|
from transformers import pipeline |
|
|
|
|
|
# Load the NER model |
|
|
ner_pipeline = pipeline('ner', model='jamshidahmadov/roberta-ner-uz', tokenizer='jamshidahmadov/roberta-ner-uz') |
|
|
|
|
|
# Example usage |
|
|
text = "Shvetsiya bosh vaziri Stefan Lyoven Stokholmdagi Spendrups kompaniyasiga tashrif buyurdi." |
|
|
entities = ner_pipeline(text) |
|
|
|
|
|
for entity in entities: |
|
|
print(entity) |
