Rapidata/Imagen-4-ultra-24-7-25_t2i_human_preference

这个数据集包含了超过83,000名个体注释者提供的近40万个关于图像比较的人类响应。数据集用于评估Imagen 4 Ultra模型在三个类别中的性能：偏好、一致性和对齐。它包括特征，如提示、图像、模型名称和每个类别的分数。数据集分为包含55,876个示例的训练集，可供下载。README还提到了用于收集数据的技术，Rapidata的Python API，并鼓励用户对数据集提供反馈，并请求未来更多的数据集。

This dataset contains nearly 400,000 human responses from over 83,000 individual annotators regarding image comparisons. It is used to evaluate the performance of the Imagen 4 Ultra model across three categories: preference, coherence, and alignment. The dataset includes features such as prompts, images, model names, and scores for each category. It is split into a training set with 55,876 examples and is available for download. The README also mentions the technology used for collecting the data, Rapidatas Python API, and encourages users to provide feedback on the dataset and request more in the future.