Fidel: A Large-Scale Sentence Level Amharic OCR Dataset
Overview
Fidel is a comprehensive dataset for Amharic Optical Character Recognition (OCR) at the sentence level. It contains a diverse collection of Amharic text images spanning handwritten, typed, and synthetic sources. This dataset aims to advance language technology for Amharic, serving critical applications such as digital ID initiatives, document digitization, and automated form processing in Ethiopia.… See the full description on the dataset page: https://huggingface.co/datasets/upanzi/fidel-dataset.
