Faizaniqbal/IndicOCR

Large-scale multilingual OCR and document dataset across 23 Pan-Indic languages and 12 writing systems. 1. Overview IndicOCR (IndicPixel) is a large-scale multilingual Optical Character Recognition (OCR) and document dataset covering the South Asian linguistic landscape. The dataset provides dense document coverage across 23 official and literary languages representing 12 distinct writing systems. Dataset Specifications: Scale & Scope: Over 12… See the full description on the dataset page: https://huggingface.co/datasets/Faizaniqbal/IndicOCR.