Misraj/Arabic-Image-Captioning_100M

这是一个包含1亿个阿拉伯语图像标题的大型多模态数据集，是首个此类规模和质量级别的阿拉伯语多模态资源。该数据集使用Mutarjim翻译模型生成，旨在填补阿拉伯语多模态AI资源的空白，并使得研究人员能够首次开发出复杂的阿拉伯语视觉-语言系统。

This is a large-scale multimodal dataset containing 100 million Arabic image captions, representing the first resource of this scale and quality in the Arabic multimodal domain. Generated using the Mutarjim translation model, this dataset addresses the critical gap in Arabic multimodal AI resources and enables researchers to develop sophisticated Arabic vision-language systems for the first time.