lmms-lab/LLaVA-ReCap-CC12M

该数据集包含图像和对话数据，每个样本包括一个唯一的ID、一张图像、一段对话（包含对话来源和内容）以及数据来源信息。数据集分为一个训练集，包含9,531,765个样本，总大小为471,267,276,831.375字节。下载大小为464,642,939,824字节。数据集的配置名为default，数据文件路径为data/train-*。

This dataset contains image and conversation data, with each sample including a unique ID, an image, a conversation (containing the source and content of the conversation), and data source information. The dataset is divided into a training set containing 9,531,765 samples, with a total size of 471,267,276,831.375 bytes. The download size is 464,642,939,824 bytes. The dataset configuration is named default, and the data file path is data/train-*.