dlp-sample

IndicDLP Sample Sample subset of the IndicDLP (Indic Document Layout Parser) dataset. 23 languages: Assamese, Bengali, Bodo, Dogri, English, Gujarati, Hindi, Kannada, Kashmiri, Konkani, Malayalam, Maithili, Manipuri, Marathi, Nepali, Odia, Punjabi, Sanskrit, Sindhi, Santali, Tamil, Telugu, Urdu 12 document domains: Acts & Rules, Brochure, Form, Magazine, Manual, Newspaper, Notice, Novel, Question Paper, Research Paper, Syllabus, Textbook 42 layout classes (bounding box annotations)… See the full description on the dataset page:

Source: Hugging Face Hub (onath/dlp-sample). Metadata imported from the dataset’s Hub tags.