DocVQA
Document Visual Question Answering benchmark focused on extracting information from scanned documents or table images. It evaluates a model's OCR capability and its ability to understand document structure.
Timeline (last 90 days)
Different datasets such as FUNSD, CORD, and DocVQA do not have unified definitions of annotation granularity and boundaries.
DocVQA focuses on extracting information from scanned documents or table images, testing a model's OCR and structural comprehension capabilities.
The Qwen2.5-VL series performs especially well on the DocVQA and OCRBench benchmarks.
All Facts (3)
Different datasets such as FUNSD, CORD, and DocVQA do not have unified definitions of annotation granularity and boundaries.
50%UnverifiedDocVQA focuses on extracting information from scanned documents or table images, testing a model's OCR and structural comprehension capabilities.
50%UnverifiedThe Qwen2.5-VL series performs especially well on the DocVQA and OCRBench benchmarks.
50%