haggingfacehyz/monology-pile-uncopyrighted-gpt2-1024
收藏Hugging Face2025-01-14 更新2025-08-30 收录
下载链接：
https://hf-mirror.com/datasets/haggingfacehyz/monology-pile-uncopyrighted-gpt2-1024
下载链接
链接失效反馈官方服务：
资源简介：
这是一个名为monology-pile-uncopyrighted的数据集，使用gpt2的Tokenizer处理，包含序列化的整数特征(input_ids)。数据集被划分为训练集(train)，包含287,567,776个示例，总大小为1,170,278,816,000字节。该数据集的上下文大小设置为1024。
This is a dataset named monology-pile-uncopyrighted, processed with gpt2 Tokenizer, containing integer sequence features (input_ids). The dataset is split into a training set (train) with 287,567,776 examples, totaling 1,170,278,816,000 bytes in size. The context size for this dataset is set to 1024.
提供机构：
haggingfacehyz
