Share your thoughts, 1 month free Claude Pro on usSee more

Home/Benchmarks

Text Extraction on FUNSD n=100 (train)

0.517Mean CER

RaV-IDP

Updated 1mo ago

Evaluation Results

Method	Links
RaV-IDP 2026.04		0.517	0.516	0.295	6	0.528

SOTA Paper

RaV-IDP

RaV-IDP: A Reconstruction-as-Validation Framework for Faithful Intelligent Document Processing

Dataset

FUNSD

Follow for update

@wizwand_team Discord

Related Benchmarks

Document Question Answering on DocVQA (val)Image Extraction and Enrichment on ScanBank (train)Table Extraction on PubTabNet n=500 (val)Text Extraction on Native PDFs (25 arXiv papers)

© 2026 wizwand

Blog Contact Changelog Swarm

Privacy Policy Terms of Service FAQs Swarm Docs