PileOfLaw-tokenized
收藏资源简介:
该数据集包含两个主要特征:'input_ids'(int32列表)和'attention_mask'(int8列表)。数据集被划分为22个分片,每个分片包含50,000个样本,大小约为1,024,400,000字节。总下载大小为10,624,906,274字节,数据集总大小为22,536,800,000字节。默认配置下,每个分片的数据文件路径已明确列出。该数据集适用于需要处理大规模序列数据的任务,如自然语言处理中的文本分类、序列标注等。
This dataset includes two core features: 'input_ids' (a list of int32 values) and 'attention_mask' (a list of int8 values). The dataset is divided into 22 shards, each containing 50,000 samples with an approximate size of 1,024,400,000 bytes. The total download size of the dataset is 10,624,906,274 bytes, and the total storage size of the full dataset is 22,536,800,000 bytes. Under the default configuration, the file paths for each shard are explicitly specified. This dataset is suitable for tasks involving large-scale sequence data processing, such as text classification and sequence labeling in natural language processing (NLP).



