skymizer/dolma-v1_7-50B-1-of-2
收藏Hugging Face2025-01-15 更新2025-02-15 收录
下载链接：
https://hf-mirror.com/datasets/skymizer/dolma-v1_7-50B-1-of-2
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了文本数据，每个样本具有一个唯一的ID和词汇计数。数据集分为训练集，总大小约为100GB，包含大约3427万条样本。提供了默认配置以访问训练数据。
The dataset consists of text data, with each sample having a unique ID and token count. The dataset is split into a training set, which is approximately 100GB in size and contains about 34.27 million samples. A default configuration is provided to access the training data.
提供机构：
skymizer
