SmallDoge/MoE_dataset
收藏Hugging Face2025-06-30 更新2025-07-05 收录
下载链接：
https://hf-mirror.com/datasets/SmallDoge/MoE_dataset
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含两个配置：默认配置和衰减配置。每个配置都只有一个训练集，包含名为input_ids的序列特征，数据类型为32位整数。默认配置的训练集大小为23.5GB，共有约143.6万个示例；衰减配置的训练集大小为62.7GB，共有约385.1万个示例。
The dataset consists of two configurations: default and decay. Each configuration has only a training set, which contains a sequence feature named input_ids with data type of 32-bit integer. The training set size of the default configuration is 23.5GB, with approximately 1.436 million examples; the training set size of the decay configuration is 62.7GB, with approximately 3.851 million examples.
提供机构：
SmallDoge
