D.Html: Document Images with Structured HTML and Markdown Markup

D.Html contains fewer than 1,000 document page images paired with structured HTML and Markdown markup for OCR and reconstruction tasks. Developed by prithivMLmods and updated in March 2026, the collection focuses on preserving document hierarchies like headings and paragraphs.

Use Cases

• Training multimodal models to map visual document features to HTML tags
• Benchmarking OCR performance on structural elements like headings and paragraphs
• Fine-tuning vision-language models for document-to-markup translation

Strengths

• Provides structured HTML/Markdown labels for layout-aware OCR
• Apache 2.0 licensed for open use
• Available in optimized Parquet format

Limitations

• Small sample size of fewer than 1,000 records
• Limited geographic or document-type diversity information

Provenance

Freshness

Last updated March 2026.

Geography

United States