Rapidata/sora-video-generation-alignment-likert-scoring

这是一个包含约6000名人类评估者对AI生成视频与给定提示匹配程度评分的数据集。评估者基于Likert量表对视频进行评分，分数越低表示视频与提示的匹配度越好。数据集包含了视频信息、用户详细信息和评分等字段，可用于视频分类和文本到视频等任务。

This dataset contains ratings from approximately 6000 human evaluators on the alignment between AI-generated videos and given prompts. Evaluators rate the videos based on the Likert scale, where a lower score indicates better alignment with the prompt. The dataset includes video information, user details, and ratings, and can be used for tasks such as video classification and text-to-video.