Nepali Synthetic OCR Lines for Historical Document Simulation

A dataset of synthetic Devanagari text line images imitating historical and official scanned document conditions. It was generated using the Mountmind PeakOCR Studio synthetic corpus generator pipeline and is designed for training OCR and HTR models. The dataset was last updated on June 11, 2026, by the author prashant0919.

Use Cases

• Training OCR models based on synthetic images of Devanagari text lines.
• Improving HTR model robustness based on simulated historical document aging artifacts.
• Benchmarking text recognition systems on realistic, degraded document conditions described in the dataset.
• Developing domain-specific recognition tools for Nepali official or historical archives.

Strengths

• Designed with realistic document aging artifacts like skew angle rotations.
• Specifically generated for Nepali Devanagari script, targeting a defined language domain.
• Intended for training modern OCR/HTR architectures like TrOCR, CRNN, and PaddleOCR.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

huggingface

Collection Method

Synthetically generated using the Mountmind PeakOCR Studio corpus generator pipeline.

Freshness

Last updated 2026-06-11 08:55:36; freshness should be verified.

Geography

Nepali language focus suggests primary relevance to Nepal.