Nayana-cognitivelab/NayanaSectionOCR

🔍 Nayana SectionOCR - Advanced Multilingual OCR Model

Developed by: CognitiveLab
License: Apache 2.0
Base Model: unsloth/gemma-3n-E4B-it
Architecture: Gemma 3n (4B parameters)

🌟 Model Overview

Nayana SectionOCR is an advanced multilingual vision-language model specifically fine-tuned for Optical Character Recognition (OCR) and Document Visual Question Answering (Document VQA) tasks. Built on the powerful Gemma 3n architecture, this model excels at understanding and extracting text from complex visual documents across multiple languages.

🌍 Supported Languages

• English (en) - Primary language
• Kannada (kn) - Indian regional language
• Hindi (hi) - Indian national language
• Marathi (mr) - Indian regional language
• Sanskrit (sa) - Classical language

17 other languages coming soon !!!!!

🎯 Key Features

• Multilingual OCR: Accurate text extraction in 5 languages
• Document Understanding: Advanced layout and structure comprehension
• Fast Inference: Optimized for real-time applications
• High Accuracy: Fine-tuned on diverse document datasets
• Easy Integration: Compatible with Transformers and Modal deployment

📋 Model Specifications

Parameter	Value
Model Size	4B parameters
Context Length	32K tokens
Image Resolution	Flexible (optimized for documents)
Precision	BFloat16
Framework	Transformers + Unsloth

🚀 Quick Start

Installation

pip install transformers torch pillow unsloth

Basic Usage

from transformers import AutoProcessor, AutoModelForImageTextToText
from PIL import Image
import torch

# Load model and processor
model_id = "Nayana-cognitivelab/NayanaSectionOCR"
processor = AutoProcessor.from_pretrained(model_id, trust_remote_code=True)
model = AutoModelForImageTextToText.from_pretrained(
    model_id,
    device_map="auto",
    torch_dtype=torch.bfloat16,
    trust_remote_code=True
)

# System prompt
system_prompt = "You are Nayana, an advanced AI assistant developed by CognitiveLab. You specialize in vision-based tasks, particularly Optical Character Recognition (OCR) and Document Visual Question Answering (Document VQA). You are highly accurate, fast, and reliable when working with complex visual documents. Most importantly, you are multilingual, capable of understanding and processing documents in a wide range of languages with precision."

# Load and process image
image = Image.open("your_document.jpg")
language = "English"  # or "Kannada", "Hindi", "Marathi", "Sanskrit"
user_prompt = f"Extract the text from this image in {language}"

# Prepare messages
messages = [
    {
        "role": "system",
        "content": [{"type": "text", "text": system_prompt}]
    },
    {
        "role": "user", 
        "content": [
            {"type": "text", "text": user_prompt},
            {"type": "image", "image": image}
        ]
    }
]

# Apply chat template
inputs = processor.apply_chat_template(
    messages,
    add_generation_prompt=True,
    tokenize=True,
    return_dict=True,
    return_tensors="pt"
)

# Generate response
with torch.inference_mode():
    outputs = model.generate(
        **inputs,
        max_new_tokens=1024,
        temperature=1.0,
        top_p=0.95,
        top_k=64,
        do_sample=True
    )

# Decode response
response = processor.tokenizer.decode(
    outputs[0][inputs["input_ids"].shape[1]:], 
    skip_special_tokens=True
)
print(response)

This model was trained 2x faster with Unsloth and Hugging Face's TRL library.

📜 Citation

@model{nayana_sectionocr_2024,
  title={Nayana SectionOCR: Multilingual Document Understanding with Gemma 3n},
  author={CognitiveLab},
  year={2024},
  url={https://huggingface.co/Nayana-cognitivelab/SectionOCR_SFT_v3_half_en_kn_hi_sa_mr_7250}
}