upanzi/fidel-dataset

Fidel是一个全面的阿姆哈拉语光学字符识别（OCR）数据集，包含来自手写、打印和合成源头的句子级别的阿姆哈拉语文本图像。该数据集旨在推动阿姆哈拉语的语言技术发展，服务于数字身份计划、文档数字化和埃塞俄比亚的自动化表格处理等关键应用。

Fidel is a comprehensive dataset for Amharic Optical Character Recognition (OCR) at the sentence level, containing text images from 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.