cvis-tmu/llamafactory-sqa3d-traces-multiimage-vqa

这是一个包含问题、视频、答案及多种评分的数据集，分为训练集和测试集两部分。数据集中的字段包括问题ID、视频信息、问题文本、正确答案、各种评分、最佳追踪索引和路径、带有格式化信息的最佳追踪、系统提示、图像标签、带有图像标签的问题以及图像列表等。

This dataset contains questions, videos, answers, and various scores, divided into training and test sets. The fields in the dataset include question ID, video information, question text, correct answer, various scores, best trace index and path, best trace with formatting information, system prompt, image tags, question with image tags, and image list.