Donut Base Finetuned Cord V2

What is Donut Base Finetuned Cord V2 ?

Donut Base Finetuned Cord V2 is a specialized AI model designed for document analysis, particularly optimized for extracting information from Indonesian receipts. It leverages advanced computer vision and natural language processing techniques to accurately identify and extract key details such as vendor names, dates, totals, and item descriptions. This model is fine-tuned for Indonesian language support, making it highly effective for processing receipts in Indonesian.

Features

• Indonesian language support: Optimized for receipts written in Indonesian.
• High accuracy: Advanced fine-tuning ensures precise extraction of relevant information.
• Multi-format handling: Works with scanned images, digital receipts, and even rotated or blurry documents.
• Comprehensive extraction: Capable of extracting vendor names, dates, totals, and item details.
• User-friendly integration: Easily integrates with existing document processing workflows.

How to use Donut Base Finetuned Cord V2 ?

1. Install the Donut library if not already installed.

   pip install donut
   

2. Import the Donut library and load the model.

   from donut import Donut
   model = Donut.load('donut-base-finetuned-cord-v2')
   

3. Load the receipt image or document for processing.

   image_path = 'path_to_your_receipt_image.jpg'
   

4. Preprocess the image (optional: rotate, crop, or enhance for better results).
5. Run inference using the loaded model.

   result = model.infer(image_path)
   

6. Extract and process the results to retrieve the desired information.

Frequently Asked Questions

What languages does Donut Base Finetuned Cord V2 support?
Donut Base Finetuned Cord V2 is primarily optimized for Indonesian receipts but can also process basic English text within documents.

Can the model handle blurry or rotated images?
Yes, the model is designed to handle rotated, blurry, or low-quality images. However, the accuracy may improve with clearer input images.

How do I access or use this model?
You can use the Donut library to load and run the model. Install the library using pip install donut and follow the usage instructions provided.