Gemma OCR

A Streamlit application that extracts text from images using Google's Gemma 3-4B model via Clarifai.

Features

• Upload images in various formats (JPG, JPEG, PNG, BMP)
• Extract text from images using Gemma 3-4B model
• Download extracted text as a file
• Responsive UI with image display options

Requirements

• Python 3.7+
• Streamlit
• Pillow
• Clarifai

Setup

1. Clone this repository
2. Install dependencies:

   pip install streamlit pillow clarifai
   

3. Set your Clarifai PAT (Personal Access Token) as an environment variable:

   export CLARIFAI_PAT="your_clarifai_pat"
   

4. Run the application:

   streamlit run ocr_app/app.py
   

Usage

1. Upload an image containing text
2. Click "Extract Text"
3. View the extracted text
4. Download the text if needed