rithikseth/

synthetic-indic-manuscript-dataset

Synthetic historical manuscript dataset generated for OCR research.

This dataset contains synthetic manuscript folios for three Indic scripts:

Devanagari Modi Sharada

Each generated manuscript image is paired with a Markdown (.md) ground-truth annotation containing the text rendered on the image.

	Dataset Statistics




	

Script Train Validation Test Total

Devanagari 85 10 5 100

Modi 85 10… See the full description on the dataset page: https://huggingface.co/datasets/rithikseth/synthetic-indic-manuscript-dataset.