OleehyO/latex-formulas-80M

该数据集是一个包含多个子集的混合数据集，每个子集都包含图像、LaTeX公式和类别信息。这些子集包括复杂度较高的公式、矩阵形式的公式、普通公式、样本公式、符号、文本混合等类型，以及英文和中文手写公式的数据。每个子集都有一个训练集，数据量从几千到几万不等。特别注意的是，手写子集的数据完全来自现有的开源作品，包括所有测试集。

The dataset consists of multiple subsets, each containing images, LaTeX formulas, and category information. These subsets include formulas of high complexity, matrix forms, ordinary formulas, sample formulas, symbols, text hybrids, as well as English and Chinese handwritten formulas. Each subset has a training set with a number of examples ranging from a few thousand to tens of thousands. An important note is that the handwritten subset of this dataset was collected entirely from existing open source works, including all test sets.