---
title: README
emoji: 🌍
colorFrom: red
colorTo: indigo
sdk: static
pinned: false
short_description: 'Nayana : Vision AI for all'
---



<div align="center">
  <h1>Nayana - Vision AI for all</h1>
  
  <a href="https://cognitivelab.in">
  <img src="https://cdn-uploads.huggingface.co/production/uploads/6442d975ad54813badc1ddf7/jjcT87PhlfwKvgWXTj3yk.png" width="80%">
  </a>  
  <h3>Enabling Vision Language Capabilites for Low resource langauges</h3>
  <p>Initiative by <a href="https://cognitivelab.in">Cognitivelab</a> </p>


</div>

## Problem Statement  

Despite advancements in vision-language AI, a significant number of the world's languages remain underserved, leaving millions without tools to process documents in their native scripts.  

**Challenges Addressed by Nayana**:  
- **Wide Language Gap**: Lack of robust OCR solutions for a large spectrum of languages, particularly low-resource and rare languages.  
- **Script Complexity**: Supporting diverse writing systems, including those with intricate scripts, cursive styles, or mixed-language content.  
- **Scalability**: Need for adaptable models that can handle real-world multilingual document processing at scale.  

Nayana is designed to tackle these challenges by fine-tuning cutting-edge OCR models for diverse languages across multiple regions, empowering users to extract actionable insights from their documents regardless of the language or script.  


## Vision  

To democratize access to **Vision-Language AI** for all communities by empowering a wide range of languages, including low-resource and underrepresented ones, with cutting-edge OCR and document understanding capabilities.  

---

## Mission  

1. **Enhance Accessibility**: Build tools that enable equitable AI solutions for diverse linguistic groups worldwide.  
2. **Expand Language Coverage**: Support a vast range of languages and scripts, breaking barriers for multilingual document processing.  
3. **Foster Collaboration**: Provide an open-source platform where developers and researchers can enhance and expand multilingual OCR capabilities.