linjieli222/ai2thor-path-tracing-qa-train-20k-v5

该数据集包含不同配置的视觉问答数据，每个样本包括一个问题、一个答案、多个选项、俯视图图片、侧视图图片、侧视图描述、场景ID、样本ID、房间类型、方位、路径总结等信息。数据集分为2点、2点最近、2点最远、3点顺序和4点顺序等不同配置，每个配置均有对应的训练集。数据集适用于视觉问答和相关的图像理解任务。

This dataset contains visual question answering data with different configurations. Each sample includes a question, an answer, multiple choices, a top-down image, side view images, sideview description, scene ID, sample ID, room type, orientation, path summary, etc. The dataset is divided into different configurations such as 2-point, 2-point nearest, 2-point furthest, 3-point order, and 4-point order, each with its corresponding training set. The dataset is suitable for visual question answering and related image understanding tasks.