yuyq96/R1-Vision-ScienceQA
收藏Hugging Face2025-02-08 更新2025-02-15 收录
下载链接：
https://hf-mirror.com/datasets/yuyq96/R1-Vision-ScienceQA
下载链接
链接失效反馈官方服务：
资源简介：
R1-Vision项目使用的数据集包括文本数据、文本渲染数据和多模态数据。文本数据来源于Bespoke-Stratos-17k，文本渲染数据也是基于Bespoke-Stratos-17k经过格式化和渲染处理得到的，而多模态数据则来自AI2D、ScienceQA和PixMo-Cap-QA等数据集。这些数据集被用来训练一个能够处理文本和图像的推理模型。
The dataset used in the R1-Vision project includes text data, text rendering data, and multimodal data. The text data is sourced from Bespoke-Stratos-17k, the text rendering data is based on Bespoke-Stratos-17k after formatting and rendering processing, and the multimodal data comes from datasets such as AI2D, ScienceQA, and PixMo-Cap-QA. These datasets are used to train a reasoning model that can handle both text and images.
提供机构：
yuyq96
