---
library_name: transformers
tags:
- sentiment-analysis
- text-classification
- urdu
- roman-urdu
license: mit
language:
- ur
base_model:
- FacebookAI/xlm-roberta-base
---
license: mit
datasets:
- custom
metrics:
- accuracy
- f1
model-index:
- name: XLM-RoBERTa Base Roman Urdu + Urdu Sentiment
  results:
  - task:
      type: text-classification
      name: Sentiment Analysis
    dataset:
      name: Custom Urdu + Roman Urdu Dataset
      type: text
    metrics:
    - name: Validation Loss
      type: loss
      value: 0.3735
    - name: Training Loss
      type: loss
      value: 0.4441
---

# 🌍 XLM-RoBERTa Base — Urdu + Roman Urdu Sentiment Analysis

This model is fine-tuned from [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) for **sentiment classification** in **Urdu** and **Roman Urdu**.  
It predicts **Positive, Negative, and Neutral** sentiments.

---

## 📊 Training Details
- **Base Model:** xlm-roberta-base  
- **Dataset:** Custom Urdu + Roman Urdu dataset  
- **Classes:** Positive, Negative, Neutral  
- **Epochs:** 3  
- **Training Loss:** 0.4441  
- **Validation Loss (best):** 0.3735  

---

## 🚀 Usage

```python
from transformers import pipeline

classifier = pipeline("sentiment-analysis", model="tahamueed/xlm-roberta-roman-urdu-sentiment")

print(classifier("yeh movie bohot zabardast hai!"))
# [{'label': 'Positive', 'score': 0.92}]