ThatsGroes/synthetic-from-text-matching-long-tasks-danish
收藏Hugging Face2025-01-31 更新2025-02-15 收录
下载链接：
https://hf-mirror.com/datasets/ThatsGroes/synthetic-from-text-matching-long-tasks-danish
下载链接
链接失效反馈官方服务：
资源简介：
该数据集旨在为丹麦语文本匹配任务预训练或后训练嵌入模型。它包含100,000个由gemma-2-27b-it生成的样本，每个样本都包括提示（prompt）和模型输出（response）。提示分为内容（content）和角色（role）。数据集的生成基于从https://huggingface.co/datasets/ThatsGroes/text-matching-long-tasks-processed随机抽取的种子任务，并遵循了一篇特定论文中描述的数据生成过程。
This dataset is designed for pre-training or post-training embedding models for Danish text matching tasks. It consists of 100,000 samples generated with gemma-2-27b-it, each including a prompt and the models output (response). The prompt is divided into content and role. The dataset was generated based on seed tasks randomly sampled from https://huggingface.co/datasets/ThatsGroes/text-matching-long-tasks-processed, following the data generation process described in a specific paper.
提供机构：
ThatsGroes
