AIbase
EN
Home
Finetune Donut Cord V2.5
F

Finetune Donut Cord V2.5

Developed by fahmiaziz
This is a vision-language model based on the Donut architecture, specifically fine-tuned for the CORD-V2 dataset for document image-to-text tasks.
Image-to-Text
Transformers
Open Source License:Openrail#Document Image to Text#High-Precision OCR#Structured Information Extraction
Downloads 97
Release Time : 9/12/2023

Model Overview

The model can extract structured text information from document images, particularly suitable for automatic recognition and conversion of receipts, forms, and other documents.

Model Features

High Accuracy
Achieves 90% accuracy on the CORD-V2 dataset
Document Understanding
Optimized for document images, capable of handling complex document layouts
End-to-End Processing
Directly processes from image input to structured text output without intermediate steps

Model Capabilities

Document Image Recognition
Text Extraction
Structured Data Conversion
Receipt Information Extraction

Use Cases

Document Digitization
Receipt Processing
Automatically extracts merchant, date, amount, and other information from receipt images
90% accuracy
Form Recognition
Converts paper forms into structured electronic data
Office Automation
Document Archiving
Automatically generates searchable text content for scanned documents
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
English简体中文繁體中文にほんご
© 2025AIbase