abshetty/rlaif-10000

该数据集包含10000个样本，主要用于训练。每个样本包含以下字段：数据集名称（ds_name）、图像（image）、问题（question）、选择的答案（chosen）、拒绝的答案（rejected）、原始数据集名称（origin_dataset）、原始数据集分割（origin_split）、索引（idx）和图像路径（image_path）。数据集的总大小为1634651824.0字节，下载大小为1618877390字节。

This dataset contains 10,000 samples primarily intended for training. Each sample includes the following fields: dataset name (ds_name), image, question, chosen answer (chosen), rejected answer (rejected), original dataset name (origin_dataset), original dataset split (origin_split), index (idx), and image path (image_path). The total size of the dataset is 1634651824.0 bytes, with a download size of 1618877390 bytes.