phxember/Uni-MuMER-Data
收藏Hugging Face2025-09-12 更新2025-10-25 收录
下载链接：
https://hf-mirror.com/datasets/phxember/Uni-MuMER-Data
下载链接
链接失效反馈官方服务：
资源简介：
Uni-MuMER数据集包括CROHME、HME100K、MathWriting和Im2LaTeXv2等多个子集，用于手写数学表达式识别（HMER）的多任务训练。数据集包含图像和LaTeX格式的标签，以及用于结构推理、错误学习和符号计数的特定子集。数据集经过了预处理，包括语法清理、分词和归一化。数据集的大小在1M到10M之间，许可证为MIT。
The Uni-MuMER dataset includes subsets such as CROHME, HME100K, MathWriting, and Im2LaTeXv2, designed for multi-task training in Handwritten Mathematical Expression Recognition (HMER). The dataset consists of images and LaTeX format labels, with specific subsets for structural reasoning, error learning, and symbol counting. The data has undergone preprocessing, including syntax cleaning, tokenization, and normalization. The dataset size ranges between 1M and 10M, and the license is MIT.
提供机构：
phxember
