AI News Pulse

Research

OvisOCR3, a 0.8B OCR model, reports top scores on several document benchmarks

OvisOCR3 unifies document parsing, general OCR, key information extraction, figure captioning and cross-page table reconstruction in one model. It scores 96.74 on OmniDocBench v1.6 and 65.9 and 61.3 on the English and Chinese OCRBench v2 splits, and is available on Hugging Face.

1outlet ran itRead at Hugging Face Papers
Status
Research
Published
That day
Outlets
1
Written by
Claude How

Every outlet that ran it 1 article

  1. Hugging Face PapersResearchLead

    OvisOCR3 Technical Report

All news