Rapidata/text-2-video-human-preferences-moonvalley-marey

这是一个名为Moonvalley Marey Human Preferences的数据集，主要用于评估Marey视频生成模型。数据集包含了约75k份来自约20k个人类标注者的响应，用于评估Marey视频生成模型。数据集通过Rapidata Python API收集，适合大规模数据标注。数据集包含视频对比，每个条目包括video1和video2字段，包含链接到降级GIF的图片。weighted_results列包含从0到1的分数，代表聚合的用户响应。detailedResults列包含单个用户的响应。数据集还包含关于Alignment、Coherence和Preference的评分，分别表示视频与提示的匹配程度、视频的连贯性和视觉吸引力。

This is a dataset named Moonvalley Marey Human Preferences, mainly used to evaluate the Marey video generation model. The dataset contains about 75k responses from about 20k human annotators for evaluating the Marey video generation model. The dataset was collected through the Rapidata Python API, which is suitable for large-scale data annotation. The dataset includes video comparisons, with each entry including video1 and video2 fields, containing links to downsampled GIF images. The weighted_results column contains scores ranging from 0 to 1, representing aggregated user responses. The detailedResults column contains individual user responses. The dataset also includes scores for Alignment, Coherence, and Preference, representing the degree of match between the video and the prompt, the coherence of the video, and the visual appeal of the video, respectively.