Unsplash LITE Dataset

该数据集是从Unsplash图片库中筛选出的一组图片，包含了25,000张可供商业用途的图片，用于研究当创作者的作品被用作人工智能训练数据时，他们的行为反应。此外，该数据集还包括了诸如上传日期、用户信息、关键词以及相机设备细节等元数据信息，同时允许追踪用户的活动以及随时间变化的投稿动态。在Unsplash提供的总共约490万张图片中，这个数据集包含了大约25,000张图片。该数据集的任务是分析创作者在面对作品被用作人工智能训练数据时，其投稿行为及上传动态的变化。

This dataset comprises approximately 25,000 commercially usable images curated from the Unsplash image library, and is developed to study creators' behavioral responses when their works are utilized as AI training data. In addition, the dataset includes comprehensive metadata such as upload dates, user information, keywords, and camera device details, enabling tracking of user activities and temporal submission trends. This is a curated subset selected from the total pool of around 4.9 million images hosted on Unsplash. The core research task of this dataset is to analyze shifts in creators' submission behaviors and upload dynamics when their works are employed for AI training purposes.