Im2latex-90k

由 Singh、Sumeet S. 介绍。“教学机器编码：具有视觉注意的神经标记生成”。 ArXiv abs/1802.05415 (2018): n。页。 im2latex-100K 数据集的净化版本（删除了错误的样本）。用于 OpenAI 的 image-2-latex 系统任务的预构建数据集。包括总共约 90k 的公式和图像，分为训练集、验证集和测试集。另请参阅 I2L-140K，它是该数据集的超集。Introduced by Singh, Sumeet S. in "Teaching Machines to Code: Neural Markup Generation with Visual Attention". ArXiv abs/1802.05415 (2018): n. pag. This is a cleaned version of the im2latex-100K dataset, with erroneous samples removed. It is a pre-built dataset for the OpenAI image-to-latex system task, containing approximately 90k total formulas and images split into training, validation, and test sets. Also refer to I2L-140K, which is a superset of this dataset.

搜集汇总

背景与挑战

背景概述

Im2latex-90k是im2latex-100K数据集的净化版本，专门用于OpenAI的image-2-latex系统任务，包含约9万个公式和图像，并划分为训练、验证和测试集。该数据集的超集为I2L-140K。