MEPAVE

该数据集是一个多模态的数据集，包含了产品图片以及相关的中文文本信息。在本研究中，我们仅使用中文文本信息来对模型进行训练和评估。该数据集用于评估所提出模型预测新值的能力，并与其他强基线方法进行比较，使用的评估指标包括F1分数和精确匹配度。这项任务的目标是进行属性值提取。

This is a multimodal dataset containing product images and their associated Chinese textual information. In this study, we exclusively use the Chinese textual information for model training and evaluation. This dataset is employed to evaluate the capability of the proposed model to predict new values, and to compare it with other strong baseline methods, with evaluation metrics including F1 score and Exact Match. The objective of this task is attribute value extraction.