mlfoundations-dev/original_tiger_dataset_full

该数据集包含问题、答案和对话三个主要特征。对话部分由来源和值组成。数据集仅包含一个训练集，样本数量为11,621,594个，文件大小为22,181,152,326字节。下载大小为11,278,360,159字节。数据集的配置名为default，数据文件路径为data/train-*。

The dataset contains three main features: question, answer, and conversations. The conversations feature is a list containing from and value. The dataset includes only a training set with 11,621,594 examples and a file size of 22,181,152,326 bytes. The download size is 11,278,360,159 bytes. The dataset configuration is named default, and the data files are located at data/train-*.