3xartis7/laion-1m

LAION-Subset (Lance Format) 是一个基于Lance格式的LAION图像-文本语料库子集，包含约100万行数据。数据集提供内联JPEG字节、CLIP嵌入（img_emb）和完整元数据，可直接从Hub访问。它适用于特征提取和图像分类任务，主要用于多模态AI工作负载，支持快速随机访问、原生多模态数据存储和高效查询。LAION-Subset (Lance Format) is a subset of the LAION image-text corpus in Lance format, containing approximately 1 million rows of data. The dataset provides inline JPEG bytes, CLIP embeddings (img_emb), and full metadata, and is directly accessible via the Hub. It is suitable for feature extraction and image classification tasks, primarily used for multimodal AI workloads, and supports fast random access, native multimodal data storage, and efficient querying.