Vokturz/sourceforge-app-screenshots-ocr

该数据集包含了来自SourceForge的应用程序屏幕截图，以及由特定模型生成的元数据和OCR文本。数据集用于为Loyca-ai项目微调更小的Qwen3-VL模型。数据集的特征包括唯一标识符、图像描述、关键词、类别、OCR文本以及屏幕截图本身。数据集被分为训练集，并且提供了下载和数据集的大小信息。

The dataset consists of screenshots from SourceForge applications, along with metadata and OCR text generated by a specific model. It is used to fine-tune smaller Qwen3-VL models for the Loyca-ai project. The dataset features include unique identifiers, image descriptions, keywords, categories, OCR text, and the screenshots themselves. The dataset is split into a training set, and the sizes for download and the dataset are provided.