rookshanks/pile_uncopyrighted_pythia_1024
收藏Hugging Face2025-05-28 更新2025-10-25 收录
下载链接：
https://hf-mirror.com/datasets/rookshanks/pile_uncopyrighted_pythia_1024
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了三个字段：input_ids（整数序列），attention_mask（字节序列），length（整数）。数据集分为训练集、测试集和验证集，其中训练集包含47776288个示例，测试集包含20120个示例，验证集包含2240个示例。数据集的总大小为169980225748字节，下载大小为67753861127字节。
The dataset includes three fields: input_ids (integer sequence), attention_mask (byte sequence), and length (integer). It is divided into training, test, and validation sets, with the training set containing 47776288 examples, the test set containing 20120 examples, and the validation set containing 2240 examples. The total size of the dataset is 169980225748 bytes, and the download size is 67753861127 bytes.
提供机构：
rookshanks
