issai/BabyVision_Kazakh

BabyVision-Kazakh是BabyVision数据集的机器翻译版本，专门用于评估多模态大语言模型的类人视觉感知能力。它专注于对年轻儿童直观但对先进AI系统仍具挑战性的任务，如基本计数、形状识别和异常识别。数据集包含图像、问题、选项和答案等字段，任务类型为视觉问答。

BabyVision-Kazakh is a machine-translated version of the BabyVision dataset, designed to evaluate the human-like visual perception of Multimodal Large Language Models. It focuses on tasks that are intuitive for young children but remain challenging for advanced AI systems, such as basic counting, shape recognition, and identifying anomalies. The dataset includes fields like image, question, options, and answer, with the task category being visual-question-answering.