README.md

---
{}
---
library_name: transformers
tags:
  - amharic
  - ner
  - token-classification
  - xlm-roberta
---

# Ethio NER Model

Fine-tuned Amharic Named Entity Recognition (NER) model based on XLM-Roberta. Built for extracting entities from Telegram-based e-commerce messages.

## Usage

```python
from transformers import AutoTokenizer, AutoModelForTokenClassification, pipeline

tokenizer = AutoTokenizer.from_pretrained("rufeshe/ethio-ner-model")
model = AutoModelForTokenClassification.from_pretrained("rufeshe/ethio-ner-model")

nlp = pipeline("ner", model=model, tokenizer=tokenizer, aggregation_strategy="simple")

text = "አዋስማርት በአዲስ ሲስተም ላይ አዳዲስ ሞዴሎችን አመጣ።"
print(nlp(text))
Training Details
Base model: xlm-roberta-base

Data: Cleaned Telegram messages from Ethiopian e-commerce channels

Framework: 🤗 Transformers

Metrics: F1 score, Precision, Recall

Model Card
Developed by: [Your Name]
Shared by: [Your Organization]
License: Apache-2.0
Language: Amharic (am)

Contact
For feedback or questions, reach out at zerufeshetu121@gmail.com