allenai/olmo-mix-1124
收藏Hugging Face2025-08-19 更新2024-12-14 收录
下载链接:
https://hf-mirror.com/datasets/allenai/olmo-mix-1124
下载链接
链接失效反馈官方服务:
资源简介:
OLMo 2 Mix (November 2024)数据集是一个用于训练OLMo-2-1124模型的数据集,主要来源于DCLM-Baseline,未进行额外过滤。数据集包含多个组成部分,如Arxiv、pes2o、starcoder、Algebraic-stack、OpenWebMath和Wiki等,每个部分都有详细的令牌数、字节数、文档数及许可证信息。整个数据集的总令牌数为3.90T,总字节数为22.4TB,总文档数为3.08M,许可证为ODC-BY。
The OLMo 2 Mix (November 2024) dataset is a collection of data used to train OLMo-2-1124 models, primarily sourced from DCLM-Baseline with no additional filtering. The dataset includes multiple components such as Arxiv, pes2o, starcoder, Algebraic-stack, OpenWebMath, and Wiki, each with detailed token counts, byte sizes, document counts, and license information. The entire dataset totals 3.90T tokens, 22.4TB in size, and 3.08M documents, licensed under ODC-BY.
提供机构:
allenai



