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

这是一个使用t2v Sora风格的Likert评分的视频生成偏好数据集。数据集中大约有6000名人类评估者被要求根据视频的视觉吸引力对AI生成的视频进行评分，而不看到用于生成视频的提示。评分使用五级Likert量表，评估者的详细个人信息也包含在数据集中。

This is a t2v Sora Style Likert Scores video generation preference dataset. In this dataset, approximately 6000 human evaluators were asked to rate AI-generated videos based on their visual appeal without seeing the prompts used to generate them. The ratings were made using a five-point Likert scale, and detailed personal information of the evaluators is also included in the dataset.