遇见数据集

shiwk24/MathCanvas-Imagen

收藏
Hugging Face2025-11-18 更新2025-10-25 收录
官方服务:

资源简介:

MathCanvas-Imagen数据集是一个大规模的数据集,包含超过1000万对文本描述与对应数学图表的配对,是MathCanvas框架的核心组成部分。该数据集旨在训练多模态模型解决复杂数学问题的内在视觉链式思维能力。数据集由五个不同的子集组成,涵盖了从竞赛级复杂数学问题到基础几何代数结构的各种图表和描述。

MathCanvas-Imagen is a massive dataset with over 10 million caption-to-diagram pairs, forming a core part of the MathCanvas framework. It is designed to train Unified Large Multimodal Models (LMMs) with intrinsic Visual Chain-of-Thought (VCoT) capabilities for solving complex mathematical problems. The dataset consists of five distinct subsets, covering a range from complex competitive math problems to basic geometric and algebraic structures.

提供机构:
shiwk24
二维码
社区交流群
二维码
科研交流群
商业服务