What it is
Benchmark for instruction-guided editing of dense English and Chinese document images, with localized text/table changes and reference outputs. 3,506 document-image editing examples with input page, edit instruction, target page, and region/change masks across 7 text and table edit types in English and Chinese.
Key results
- 3,506 test examples total
- 7 edit types: 3 text and 4 table tasks
- Bilingual English and Chinese document images
Useful for benchmarking models that must edit document images locally while preserving layout, typography, and nearby content.