tungvu3196/vlm-projects-multi-lang-final-2

该数据集包含多个语言配置，每个配置都包含相同的特征，如A1至A4、归一化的边界框坐标、可交付成果、备注、患者ID、Q1至Q4、幻灯片以及带和不带边界框的图像。数据集分为训练集和测试集，每个子集都有指定数量的示例和数据大小。数据集支持阿拉伯语、英语、法语、德语、印地语、印度尼西亚语、日语、韩语、马来语、普通话、葡萄牙语、俄语、西班牙语、泰语、土耳其语和越南语。每种语言配置都有自己的训练和测试数据文件路径。

The dataset contains multiple language configurations, each including the same set of features such as A1 to A4, normalized bounding box coordinates, deliverable, notes, patient ID, Q1 to Q4, slide, and images with and without bounding boxes. The dataset is split into train and test subsets, with specified number of examples and byte sizes for each. The dataset is available in Arabic, English, French, German, Hindi, Indonesian, Japanese, Korean, Malay, Mandarin, Portuguese, Russian, Spanish, Thai, Turkish, and Vietnamese languages. Each language configuration has its own data files path for train and test splits.