HuggingFaceTB/dclm-edu
收藏Hugging Face2025-03-07 更新2025-08-30 收录
下载链接：
https://hf-mirror.com/datasets/HuggingFaceTB/dclm-edu
下载链接
链接失效反馈官方服务：
资源简介：
DCLM-Edu是一个经过筛选的DCLM数据集版本，使用FineWeb-Edu教育质量分类器对网页进行标注，并只包含教育质量评分高于2的样本。该数据集旨在用于小语言模型的训练。
DCLM-Edu is a filtered version of the DCLM dataset, annotated with educational quality using the FineWeb-Edu classifier, and only includes samples with an educational quality score higher than 2. This dataset is intended for training small language models.
提供机构：
HuggingFaceTB
