ibm-esa-geospatial/Llama3-SSL4EO-S12-v1.1-captions

Llama3-SSL4EO-S12-Captions数据集包含了与SSL4EO-S12 v1.1卫星图像数据集对齐的字幕，这些字幕是使用Llama3-LLaVA-Next-8B模型自动生成的。数据集分为训练集和验证集，提供CSV和Parquet两种文件格式。每个样本包括图像字幕、相关的问题-答案对、图像文件路径和样本ID等信息。The Llama3-SSL4EO-S12-Captions dataset contains captions aligned with the SSL4EO-S12 v1.1 satellite imagery dataset, automatically generated using the Llama3-LLaVA-Next-8B model. The dataset is split into training and validation sets and provides files in both CSV and Parquet formats. Each sample includes an image caption, associated question-answer pairs, the path to the image file, and the sample ID.