ahmedheakl/ar_vqav2_instruct

该数据集包含多个特征，如问题ID、图像ID、问题、问题类型、答案类型、多项选择答案、图像路径等。此外，数据集还包含了多个答案及其置信度和ID，以及图像和对话内容。数据集分为训练集，包含50,000个样本，总大小为7,859,267,723字节。The dataset contains multiple features such as question_id, image_id, question, question_type, answer_type, multiple_choice_answer, image_path, etc. Additionally, the dataset includes multiple answers along with their confidence levels and IDs, as well as images and conversation content. The dataset is divided into a training set containing 50,000 samples, with a total size of 7,859,267,723 bytes.