Skip to content

Dataset · hyperbots-inc

Document data extraction dataset

hyperbots-inc/synthetic-enterprise-doc-extraction

372 synthetic enterprise documents of various types with JSON templates for benchmarking.

What's inside

The set includes 372 synthetic documents of 31 types (invoices, bills, KYC forms, medical leaves, etc.). Each record contains ground truth field values and a JSON template for data extraction. Data is in English, volume is less than 1K examples.

What you can do with it

  • Benchmarking structured document data extraction systems
  • Testing schema-oriented document processing pipelines
Tasks
Text generation, Document question answering
Languages
English
Size
up to 1K rows
Downloads this month
259
Likes
44

How to load it

from datasets import load_dataset

ds = load_dataset("hyperbots-inc/synthetic-enterprise-doc-extraction")