Repository files navigation

A powerful OCR (Optical Character Recognition) package that uses state-of-the-art vision language models through Ollama to extract text from images and PDF. Available both as a Python package and a Streamlit web application.

• Multiple Vision Models Support

  • LLaVA: Efficient vision-language model for real-time processing (LLaVa model can generate wrong output sometimes)
  • Llama 3.2 Vision: Advanced model with high accuracy for complex documents
  • Granite3.2-vision: A compact and efficient vision-language model, specifically designed for visual document understanding, enabling automated content extraction from tables, charts, infographics, plots, diagrams, and more.
  • Moondream: Small vision language model designed to run efficiently on edge devices.
  • Minicpm-v: MiniCPM-V 2.6 can process images with any aspect ratio and up to 1.8 million pixels (e.g., 1344x1344).

• Multiple Output Formats

  • Markdown: Preserves text formatting with headers and lists
  • Plain Text: Clean, simple text extraction
  • JSON: Structured data format
  • Structured: Tables and organized data
  • Key-Value Pairs: Extracts labeled information
  • Table: Extract all tabular data.

• Batch Processing

  • Process multiple images in parallel
  • Progress tracking for each image
  • Image preprocessing (resize, normalize, etc.)

• Custom Prompts

  • Override default prompts with custom instructions for text extraction.

pip install ollama-ocr

1. Install Ollama
2. Pull the required model:

ollama pull llama3.2-vision:11bollama pull granite3.2-visionollama pull moondreamollama pull minicpm-v

fromollama_ocrimportOCRProcessor# Initialize OCR processorocr=OCRProcessor(model_name='llama3.2-vision:11b',base_url="http://host.docker.internal:11434/api/generate")# You can use any vision model available on Ollama# you can pass your custom ollama api# Process an imageresult=ocr.process_image(image_path="path/to/your/image.png",# path to your pdf files "path/to/your/file.pdf"format_type="markdown",# Options: markdown, text, json, structured, key_valuecustom_prompt="Extract all text, focusing on dates and names.",# Optional custom promptlanguage="English"# Specify the language of the text (New! 🆕))print(result)

fromollama_ocrimportOCRProcessor# Initialize OCR processorocr=OCRProcessor(model_name='llama3.2-vision:11b',max_workers=4)# max workers for parallel processing# Process multiple images# Process multiple images with progress trackingbatch_results=ocr.process_batch(input_path="path/to/images/folder",# Directory or list of image pathsformat_type="markdown",recursive=True,# Search subdirectoriespreprocess=True,# Enable image preprocessingcustom_prompt="Extract all text, focusing on dates and names.",# Optional custom promptlanguage="English"# Specify the language of the text (New! 🆕))# Access resultsforfile_path,textinbatch_results['results'].items():print(f"\nFile:{file_path}")print(f"Extracted Text:{text}")# View statisticsprint("\nProcessing Statistics:")print(f"Total images:{batch_results['statistics']['total']}")print(f"Successfully processed:{batch_results['statistics']['successful']}")print(f"Failed:{batch_results['statistics']['failed']}")

1. Markdown Format: The output is a markdown string containing the extracted text from the image.
2. Text Format: The output is a plain text string containing the extracted text from the image.
3. JSON Format: The output is a JSON object containing the extracted text from the image.
4. Structured Format: The output is a structured object containing the extracted text from the image.
5. Key-Value Format: The output is a dictionary containing the extracted text from the image.
6. Table Format: Extract all tabular data.

• User-Friendly Interface
  • Drag-and-drop file upload
  • Real-time processing
  • Download extracted text
  • Image preview with details
  • Responsive design
  • Language Selection: Specify the language for better OCR accuracy. (New! 🆕)

1. Clone the repository:

git clone https://github.com/imanoop7/Ollama-OCR.gitcd Ollama-OCR

2. Install dependencies:

pip install -r requirements.txt

2. Go to the directory where app.py is located:

cd src/ollama_ocr

3. Run the Streamlit app:

streamlit run app.py

• Ollama OCR on Colab: How to use Ollama-OCR on Google Colab.
• Example Notebook: Example usage of Ollama OCR.
• Ollama OCR with Autogen: Use Ollama-OCR with autogen.
• Ollama OCR with LangGraph: Use Ollama-OCR with LangGraph.

This project is licensed under the MIT License - see the LICENSE file for details.

Built with OllamaPowered by Vision Models