five

dogtooth/uf_Llama-3.1-Tulu-3-8B-SFT_6_Skywork-Reward-Gemma-2-27B-v0.2_weighted_sft

收藏
Hugging Face2024-12-18 更新2024-12-21 收录
下载链接:
https://hf-mirror.com/datasets/dogtooth/uf_Llama-3.1-Tulu-3-8B-SFT_6_Skywork-Reward-Gemma-2-27B-v0.2_weighted_sft
下载链接
链接失效反馈
官方服务:
资源简介:
该数据集包含三个主要特征:prompt(提示)、response(响应)和score(评分)。其中,prompt和response为字符串类型,score为浮点数类型。数据集分为一个名为train_weighted_sft的分割,包含122,270个样本,总大小为195,203,978字节。下载大小为79,547,610字节。默认配置下的数据文件路径为data/train_weighted_sft-*。

The dataset contains three main features: prompt, response, and score. The prompt and response are of string type, while the score is of float64 type. The dataset is divided into a split named train_weighted_sft, containing 122,270 samples with a total size of 195,203,978 bytes. The download size is 79,547,610 bytes. The default configuration specifies the data file path as data/train_weighted_sft-*.
提供机构:
dogtooth
二维码
社区交流群
二维码
科研交流群
商业服务